1. Introduction
Foundations are generally designed according to pressure loads acting in the vertical direction. However, many foundations are exposed to uplift loads due to both the type of load acting on them and structural reasons, such as with towers, offshore structures, traffic signs, buried pipelines, and tunnels [
1,
2,
3,
4]. The most important factor affecting the design of foundations is the behavior of the soil in which they are located. While soils have a certain capacity under pressure loads, their tensile strength is negligible. In order to absorb the tensile forces from the structures by the foundation, it is the most practical method to use the soil weight on the foundation by burying it in the soil. However, sometimes it is impractical to bury the foundation too deeply under extreme uplift loads, especially in soils where the groundwater level is high. In this case, the need arises to increase the uplift capacity of the foundation by improving the soil. In the literature, foundations are defined as anchors, and there are many studies on uplift behavior. These studies are usually in the form of small-scale laboratory studies or numerical studies on unreinforced soil.
Dickin and Leung [
5] presented the centrifugal test results to investigate the uplift behavior of anchor plates. In the study, obtained test results were compared with theoretical methods.
Chattopadhyay and Pise [
6] proposed a theoretical model for determining the break-out resistance of horizontal plate anchors. The obtained results showed that the Foster proposed method can predict the break-out factors for a wide range of values of the shear resistance angle of the sand.
In the experimental study carried out by Dickin [
7], the uplift capacity of the anchors was investigated using a centrifugal test device. In the study, the effects of anchor geometry, embedment depth, and sand density parameters on the uplift capacity were investigated. According to the results of the tests, it was observed that the uplift capacity increased depending on the increase in anchor size, embedment depth, and the density of the sand.
An expression estimating the pull-out capacity of a horizontal strip plate was developed by Frydman and Shaham [
8]. The expression was obtained using the experimental work performed within the scope of the study and the experimental results available in the literature.
In the experimental study carried out by Dickin and Leung [
9] using a centrifuge, the uplift capacity of sand-embedded enlarged-base piles was investigated. In the study, the effects of burial depth, base diameter, and sand density were examined. According to the results obtained, the uplift capacity in loose sand was lower than the results obtained from the studies available in the literature, while the values obtained for compact sand were found to be compatible with the results of the previous study.
Dickin and Leung [
10] performed centrifugal model tests to investigate the uplift capacity of belled piers in sand. According to the test results, an empirical design method was proposed for belled piers using the parameters of the foundation geometry.
A theoretical and experimental study was carried out by Ghaly and Clemence [
11] to investigate the pull-out behavior of inclined helical screw anchors. Calculation of pull-out capacity is expressed theoretically in terms of relative depth and inclination angle. In the study, the theoretical and experimental results were compared and it was seen that they were in good agreement.
Dickin and Laman [
12] compared the lifting capacity values of the strip anchor plates obtained using centrifuge experiments with the numerical analysis results using the finite element method. In the study, it was observed that the experimental and numerical results were in very good agreement until the anchor embedding ratio was 6, while there was some difference in the embedding ratios greater than 6.
In the experimental study carried out by Bildik and Laman [
13] using small-scale laboratory tests, the uplift capacity behavior of anchor plates of different sizes and geometries buried in sand was investigated. In addition, the effects of sand density and different embedding rates on the uplift capacity behavior were investigated. From the experimental results, it was determined that the density of the sand soil and the embedding rate of the anchor plate significantly affect the uplift capacity. Comparing the obtained results with the existing theoretical methods, it has been seen that the values obtained in the case of dense sand are more consistent than in the case of loose sand.
In the study by Wang et al. [
14], a finite element approach was developed to simulate the keying process of anchor plates embedded in normal consolidated clay. According to the results obtained from the study, it is stated that the loading eccentricity should not be less than half of the anchor width for a correct design in anchor keying.
Bera [
15] carried out an experimental research on the uplift capacity of anchors. In the study, model tests were carried out in order to examine the effects of parameters such as the density of the sand and the embedment depth of the anchor. According to the results obtained from the experiments, it was observed that the uplift capacity of the anchor increased depending on the increase in the embedment depth. Using the test results, a nonlinear power model was developed that gives the uplift capacity of the anchor.
Niroumand and Kassim [
16] carried out experimental and numerical studies to determine the uplift capacity of circular anchor plates. At the end of the study, it was found that the uplift capacity behavior of the circular anchor plates was experimentally and numerically compatible, and it was stated that the numerical results gave higher values in the loose sand condition than the experimental results.
Model experiments were carried out by Niroumand and Kassim [
17] to examine the uplift behavior of rectangular anchor plates. In addition, the model experimental setup was numerically modeled with the finite element method and it was seen that the results obtained were compatible with each other. In the analyses performed using the finite element method, the sand soil was modeled with the Hardening Soil model. The research showed that the finite element results for the dense and loose cases of sand soil are higher than the experimental findings.
Zhu et al. [
18] carried out a study including laboratory and field experiments to examine the uplift capacity of a new umbrella-shaped anchor type embedded in clay soil. According to the results of the experiments, they stated that the umbrella-shaped anchor has a higher uplift capacity than the conventional anchors.
Schiavon et al. [
19] investigated the behavior of a single helical anchor plate embedded in sand under uniaxial tensile monotonic and cyclic loadings by centrifuge tests. The results contribute to clarify the impact of the number of cycles and load amplitude on anchor performance.
These studies show that the most important factors affecting the uplift capacity of the anchor plates are the dimensions of the anchor plate, the depth of burial, and the density of the soil. However, in some cases, increasing the anchor size and burial depth is not economical and feasible as it increases both the amount of excavation and the amount of backfill.
Various methods are used to increase the uplift capacity performance of anchor plates. One of these methods is the use of geosynthetics, which have become widespread, especially in recent years, and contribute to the tensile capacity of the soil. Geogrids placed in the soil are used to increase the uplift capacity of the anchors. Studies on reinforced soils are more limited than unreinforced cases [
20,
21,
22,
23,
24,
25].
Krishnaswamy and Parashar [
20] investigated the uplift behavior of anchor plates in cohesive and cohesionless soils reinforced with geogrid and unreinforced cases by performing small-scale model experiments. According to the results, they stated that the uplift capacity can be increased by using geosynthetics in both cohesive and cohesionless soils.
In the study performed by Ilamparuthi and Dickin [
21], the effect of soil reinforcement on the uplift capacity of model-belled piles and piers with various geometries embedded in the sand was investigated. According to the results, the uplift response increases with geogrid cell diameter, sand density, pile bell diameter, and embedment depth.
Niroumand et al. [
22,
23] reported experimental and numerical analysis results of the uplift capacity of anchor plates with and without geogrid and grid-fixed reinforcement. At the end of the study, it was shown that using reinforcements significantly improved the uplift capacity.
Keskin [
24] carried out experimental and numerical studies on the uplift capacity of square anchor plates. In the tests, geogrids were used to reinforce the sand. The parameters investigated the effects of the depth of the geogrid, the distance between the geogrids, the reinforcement length, the burial depth, and the sand density. According to the results obtained, the uplift capacity of square plate anchors in sand can be significantly increased with geogrid reinforcement. In addition, it was stated that the burial depth and the relative density of the sand are parameters that greatly affect the uplift capacity.
Choudhary et al. [
25] investigated the uplift performance of horizontal anchor plates in geocell-reinforced sand soils with model tests. In unreinforced conditions, it has been observed that failure occurs at displacement of 3% of the anchor plate width. Additionally, in the case of using geocell and geotextile layer, the uplift capacity increased approximately 4.5 times. The optimum geocell length was found to be 5.4 times the anchor width.
When the literature is evaluated, it is seen that the experimental studies on the behavior of foundations subjected to uplift force generally cover small-scale laboratory experiments, and the number of studies using geogrid reinforcement is quite small. The fact that the uplift capacities of geogrid reinforced anchors have not been investigated by centrifuge tests stands out as an important shortcoming. The most important difference of this study from the studies in the literature is that the uplift capacity of anchor plates in reinforced soil is investigated by centrifuge tests. In the study, the effect of sand density on uplift capacity was investigated in both unreinforced and reinforced conditions. The experiments were carried out at two different burial depths and the optimum reinforcement arrangement was investigated by placing the geogrid in different geometric conditions in the tests. Finite element analyses were carried out using the prototype model of the experimental setup and the analysis results were compared with the experimental results. The findings obtained from the study showed that the burial depth and sand density significantly affect the uplift capacity, the uplift capacity of the anchor plate can be increased by using geogrid reinforcement, and the uplift capacity can be increased by placing the geogrid reinforcement into the soil at different angles.
The structural framework of this article is as follows.
Section 2 introduces the centrifuge experimental setup and presents the experimental program and the results of the experiments with and without reinforcement.
Section 3 describes the results of the finite element analysis performed using the prototype model. Comparison and discussion of the experimental and numerical results obtained are presented in
Section 4.
Section 5 draws the conclusions.
3. Finite Element Analysis
Experimental studies were modeled with using the finite element method (FEM), and the centrifuge test results were compared with the numerical results. The experiments were modeled with the PLAXIS 2D [
27] program, which can reflect the plane strain problem. There are many soil models in the finite element programs, and it is aimed to fully reflect the elastoplastic behavior of the soil properties in the study. For this purpose, the Hardening Soil model (HSM), which is one of the advanced soil models, was used to reflect the elastoplastic behavior of the soil in numerical analysis. Soil properties obtained from laboratory tests are presented in
Table 4.
In the numerical analysis, a rigid beam element of the strip anchor plate is modeled. The modulus of elasticity of the steel material was used as the stiffness values, EA and EI. Soil/structure interface behavior in PLAXIS can be modeled using parameters generated using the Rint interaction coefficient. In this study, fully rough interface conditions, Rint = 1 are assumed.
Numerical studies were carried out in three stages. In the first stage, the mesh effect was investigated and the mesh density to be used in the analysis was determined. In the second stage, the uplift capacity/displacement curves obtained from the experiments were compared, and in the last stage, the comparisons in terms of the breakout factor were presented collectively.
Investigation of Mesh Density on Results
The mesh density used in the FE program directly affects the results. There are five different mesh densities in the PLAXIS program, and the effects of five different mesh densities (very coarse, VC; coarse, C; medium, M; fine, F; and very fine, VF) on the results for the experimental model were examined before the analysis. For this purpose, the anchor plate at H/B = 5 burial depth was modeled, and the results were examined.
Table 5 presents the mesh densities used in the analysis, the number of elements for these densities, and the ultimate uplift capacities obtained as a result of the analysis.
FE models for very coarse and very fine mesh conditions for the H/B = 5 model are given in
Figure 10 and
Figure 11, respectively. The results obtained for different mesh ranges are presented in
Figure 12, and fine mesh density, which did not affect the result, was used in the analysis.