In Situ Testing of Square Footing Resting on Geobelt-Reinforced Gravel Thin Cushion on Soft Silt

A series of in situ static loading tests of square footing were carried out on the geobelt-reinforced gravel cushion on soft silt. +e reinforced gravel cushion was thin with the depth-to-width ratio of 0.2. A parameters study was conducted by considering the number of geobelt layers, the depth of the first geobelt layer beneath the footing, the vertical spacing between two geobelt layers, the linear density of reinforcement, and the material type of geobelt. +e pressure distribution on bottom of the cushion was measured. +e test results showed that the bearing capacity of reinforced gravel cushion was significantly larger than that of unreinforced gravel cushion, and the stress diffusion effect of reinforced gravel cushion was also more pronounced than that of the unreinforced cushion. +e pressure distribution on bottom of reinforced gravel cushion was in a saddle shape. According to calculation and analysis, the stress diffusion angles of reinforced cushions were all larger than 20°.


Introduction
For using it as foundation, soft silt soil has to be treated due to its low bearing capacity and large settlement.An effective treatment is to replace a shallow depth of upper soft silt with geosynthetic-reinforced gravel cushion, which is costeffective compared to other conventional methods [1,2].It has been demonstrated that geosynthetic reinforcement can improve the bearing capacity and reduce uneven settlement of the foundation [3][4][5].By conducting 65 groups of model tests, Binquet and Lee [6,7] first reported that, by reinforcing sand layer underneath the strip footing with strips of aluminum foil, the settlement and ultimate bearing capacity of the foundation was greatly improved.In literatures [8][9][10][11][12][13][14], similar results were reported by researchers.It should be noted that it is complicated to determine the ultimate bearing capacity of square footing on reinforced soil.In addition, there has been limited research investigating the stress diffusion effect of reinforced cushion.
For determining the bearing capacity and stress diffusion capacity, most studies have used small-scale models test, which would have size effect and limitations to reflect the actual behaviors of the deformation and bearing capacity of foundation [15][16][17].erefore, it is imperative to conduct full-scale in situ tests to investigate the effects of reinforcing geosynthetic materials.In the literature, geotextiles including geogrid [17][18][19], geonet [20], geocell [21,22], and fiber [23,24] are commonly used as reinforcing materials, while geobelt, a relatively new material, has been rarely utilized.Moreover, most of the previous studies have focused on the settlement and bearing capacity of foundation [1,6,9,12,15,25], while little attention has been paid to the stress diffusion angle of reinforced cushion.In practical, due to presence of geobelt, the overall rigidity and stress diffusion capacity of the reinforced cushion can be greatly improved even for thin cushions.erefore, based on the stress diffusion theory of double-layered foundation, this study aims to investigate the bearing capacity and earth pressure distribution of square footing resting on the thin gravel cushion through a series of in situ static loading tests.
e stress diffusion effect of geobelt on the reinforced gravel cushion was also studied.

Test Design and Setup.
e testing site, with a dimension of 30 m × 17 m, is located in southwest part of Taiyuan city, China.In the testing site, the natural soil beneath the gravel cushion was saturated soft silt soil.e properties of the silt soil are listed in Table 1.
Each static loading test was conducted on the eld test pit with a length of 2.3 m and width of 2.3 m within the testing site.Figure 1 shows the layout of the test setup.A steel square rigid plate with a length of 1.5 m, width of 1.5 m, and depth of 0.3 m was used as a footing to apply the loading.A gravel cushion with a dimension of 2.3 m × 2.3 m × 0.3 m with or without geobelt reinforcement was prepared before each loading test.e loading was applied by a mechanical jackbeam loading system on the gravel cushion through the square footing.
Seven dial gauges were deployed in each test, as shown in Figure 2. Four gauges were attached on the four corners of the square rigid plate to measure the overall settlement of the footing.Other three gauges were attached on the bottom of the gravel cushion to monitor the settlement of silt soil.e deformation of the gravel cushion can then be calculated according to the measurements of seven gauges.As shown in Figure 3, twenty-two pressure cells, with a measuring capacity of 0.6 MPa, were deployed on the bottom of the gravel cushion to measure the pressure of the cushion resting on silt soil during loading.
In total, 10 in situ tests were conducted on unreinforced and geobelt-reinforced gravel cushions over saturated silt soil.e testing programs are summarized in Table 2.     From construction perspective, the geobelts have the advantages of light-weight, simple construction, and shorter construction period.ere are rough details and ribs on the surface of two types of geobelt materials for improving the bonding between geobelts and gravel.eir engineering properties are listed in Table 3.
In the test, in order to ensure strong bonding between geobelts and the gravel, the geobelts are bent at both ends by using a sand bag which is around 550∼600mm and then xed by two clamps, as shown in Figure 5. erefore, the e ective length of the geobelt used in the cushion is around 3.5m, including the width of the cushion and the xing length at both ends.

Gravel Cushion.
e thin cushion consisted of gravels with diameter in the range of 10mm-30mm.Particle size distribution of the gravels is shown in Figure 6. e physical parameters obtained in laboratory tests are listed in Table 4, in which the maximum dry density and optimum moisture content were determined by Standard Proctor Test.According to USCS and AASHTO classi cation, the gravel was classi ed as poorly graded gravel (GP).

Test Procedure.
To ensure that each test was conducted under same conditions, the following preparations have to be done.First, the testing pits were dug, cleaned, and leveled to make sure their sizes meet the requirements.Secondly, the sorted ne sand was paved, compacted, and leveled on the bottom of the pits with a thickness of 10 mm∼15 mm for reducing stress concentration in the pressure cell.en the pressure cells were put on the preset positions.e pressure cells should be waterproof and calibrated before each test.Six layers of gravel were paved on the pressure cell with each layer about 50 mm thick.Each layer of gravel in the loose state had the same weight and was then compacted to same compactness by using a wood hammer.Geobelts were laid on the speci ed position in the gravel cushion with the required linear density of reinforcement (LDR) in two dimensions.e geobelts should be tightened and straightened.A layer of ne sand with a thickness of 10 mm∼20 mm was paved above and beneath the geobelts for protecting   them from not being punctured by gravel.Geobelts and gravel were laid alternately according to the requirements until the cushion height reached 300 mm.en the loading plate was put on the top of the reinforced gravel cushion, leveled, and centered properly, to ensure that the load can be evenly distributed on the gravel cushion.e static loading was applied by using a hydraulic jack.e loading method and stability criteria followed the Code for Design of Building Foundation (GB 50007-2011) [26].e loading was applied with an increment of 20 kPa.If it was less than 0.1 mm/h in two hours, the settlement was recognized to meet the criteria and the next increment of loading can be applied.e test was terminated when the total settlement reached 0.06 B, that is, 90 mm in this program.
After one test was completed, a new test pit was dug in adjacent locations within the testing site.Identical testing procedures were used to perform each test.

Reinforcing E ect on Bearing Capacity of Foundation.
Figure 7 shows pressure-settlement curves of foundations with and without geobelts.It can be seen that the settlement increased with the increase of pressure for both reinforced and unreinforced cushions.However, cushion reinforced with two geobelt layers showed the slowest increase rate of settlement, while unreinforced cushion showed the quickest increase rate.e results indicate that the bearing capacity can be e ectively improved by geobelt reinforcement.e reason is that the lateral restraint of geobelt on gravel cushion can decrease the settlement of the gravel cushion, and thus improve the bearing capacity of the foundation compared to the unreinforced cushion.At early stage, the settlement increased linearly with pressure.As pressure increased, geobelt comes into play and decreases settlement e ectively.
Since no abrupt increase of settlement was observed in all tests, the ultimate bearing capacity of foundation was determined when the settlement of foundation reached 0.06 B (B is the length of the square footing).
To account for the size e ect of footing, the bearing capacity ratio (BCR) recommended by Binquet and Lee was calculated as following: in which q R and q 0 are the bearing capacity for reinforced and unreinforced soil, respectively.For convenience, the test results were analyzed according to the BCRs calculated at di erent settlement ratios (s/B).e settlement ratio is calculated by dividing settlement of footing (s) with the width of footing (B).e values of BCR at settlement ratios (s/B) of 0.01, 0.02, 0.03, 0.04, 0.05, and 0.06 are listed in Table 5. e BCR for one-layer geobelt reinforcement was between 1.22 and 1.37, and gradually decreased with the increase of the settlement ratio; the BCR of two-layer geobelt reinforcement was between 1.34 and 1.70 and gradually increased with the increase of the settlement ratio.In current study, the values of BCR are lower than that proposed by Adams and Collin [27], and Chen et al. [28]. is may be due to the relatively small thickness of the cushion used in this study.
Figures 8(a) and 8(b) show the relationship between the number of reinforced layers (N) and BCR.As can be seen, two-layer geobelt reinforcement was much better than onelayer geobelt reinforcement, especially at the late stage of loading.At s/B of 0.06, BCRs of one-layer reinforcement were from 1.22 to 1.26, while for two-layer reinforcement, they were from 1.62 to 1.70, which indicates that the bearing  4 Advances in Materials Science and Engineering capacity of soil reinforced with two-layer geobelt can be increased more at the limit state.
Figure 9 shows the relationship between BCR and the depth of the rst geobelt layer beneath the footing (U).At the initial stage of loading (s/B ≤ 0.2), BCR reached the maximum at 50 mm (U), and gradually decreased as U increased.At the later stage of loading (s/B ≥ 0.3), BCR slightly increased and the bearing capacity of cushion was improved.is can be explained that the geobelt comes into play after certain amount of settlement happened.In addition, the geobelt could take the e ect earlier if it is closer to the footing.
Figure 10 shows the relationship between BCR and the linear density of the reinforcement (LDR).As can be seen, BCR increased slightly with the increase of LDR.When LDR increased from 25% to 50%, the value of BCR increased from (1.22∼1.29) to (1.26∼1.34).
is can be ascribed to the increase of shear strength since the interactions between the geobelt and gravel is improved when LDR is increased which provides more lateral restraint to the gravel cushion.Figures 11(a) and 11(b) demonstrate the relationship between BCR and vertical spacing (H) for two-layer geobeltreinforced cushion.At the initial stage of loading (s/B ≤ 0.2), BCR decreased with the increase of H, while at the later stage (s/B ≥ 0.3), BCR increased with the increase of H. e results also indicate that the geobelt takes e ect when certain amount of settlement happened.At the later stage of loading, reinforcing e ect of the geobelt became more pronounced as there was a continuous increase of deformation of the cushion.
Figures 12(a) and 12(b) show the bearing capacity ratio against the settlement ratio for the cushion reinforced with the TG geobelt and CPE geobelt to compare their reinforcing e ects.At the initial stage of loading, in terms of bearing  capacity, the TG-type geobelt performed better than the CPE-type geobelt.However, the di erence between them became unnoticeable with the increase of loading.

Stress Distribution at the Bottom of Reinforced Cushions.
Figure 13(a) shows the stress distribution on bottom of the unreinforced gravel cushion.Figures 13(b) and 13(c) show stress distribution on bottom of the reinforced gravel cushion under di erent pressures.As can be seen in Figure 13(a), for the unreinforced gravel cushion, the stress distribution on bottom of the cushion was parabolic with maximum net earth pressure reached at the center of the cushion.e stress distribution curves on bottom of the reinforced cushion were both in saddle shape and the maximum earth pressures were found to be about 370 mm (one-layer geobelt) and 750 mm (two-layer geobelt) away from the center with the minimum pressure at the center.For the unreinforced cushion, surrounding soft silt soil could not provide strong constrain.As loading increases, the gravels on bottom edges of the footing could be easily pushed out laterally since the pressure on bottom could not be adjusted by the unreinforced cushion itself.While for reinforced cushions, the geobelt restricted the lateral displacement of the cushion owing to frictions between the geobelt and gravels.erefore, it was not easy to push out the soils located on bottom edges of the footing.
e results indicate that, due to using the geobelt, the stress distribution was improved.e central reaction force was decreased, and the edge reaction force was increased.For one-layer and two-layer reinforced cushions (Figures 13(b) and 13(c)), it can be found that, compared to the one-layer cushion, the pressure on the edges of the two-layer reinforced cushion dramatically increased and the pressure distribution became smoother.It is clear that in terms of the reinforcing e ect, the two-layer reinforcement is much better than the onelayer reinforcement.

Stress Di usion Angle of Cushion.
As shown in Figure 13, the earth pressure measured on bottom of the cushions was not evenly distributed.So in this section, the average pressure on the bottom of cushion was calculated according to the measured settlement results which are more accurate and uniform.In same stress path, there exists a one-to-one relationship between settlement and pressure of soil layers.erefore, the average pressure (P z ) under the cushion can be obtained by using the pressure-settlement curves of natural ground without cushion and with upper cushion, as shown in Figure 14.When the settlement of natural ground under the cushion was equivalent to that of natural ground without the cushion, the corresponding relationship between the pressure P 0 under the cushion and the load P y on the top of natural ground was obtained; that is, P y is equal to the mean earth pressure of cushions P z .Based on the principle of stress di usion (Figure 15), the stress di usion angle of the reinforced cushion can be calculated by the following formula: To investigate the e ect of geobelt on stress di usion, a ratio SDR was de ned, in this study, by dividing the stress di usion angle of the geobelt reinforced cushion with that of the unreinforced cushion.Table 6 lists stress di usion angles   Current results are consistent with the result of Gabr et al. [29,30] that the stress di usion angle at the stable state was smaller than that at the proportional limit state.

Advances in Materials Science and Engineering
Figures [16][17][18][19] show the relationships between di erent parameters and the stress di usion angle ratio.As can be seen in Figures 16(a) and 16(b), under same reinforcement conditions, the two-layer geobelt-reinforced cushion showed larger stress di usion angle compared to the onelayer geobelt reinforced one.At the proportional limit state, SDR of the two-layer reinforced cushion was similar to that of the one-layer reinforced cushion, while at the stable state, there was a big di erence between them.e reason is that, for two-layer geobelt reinforcement, rst geobelt closer to the footing takes e ect when the loading is small.As the loading increases, the second geobelt would then take its e ect.erefore, at the later stage of loading, the reinforcing e ect of two-layer reinforcement became more pronounced.As can be seen in Table 6, at the stable state, SDRs spanned from 1.76 to 1.84 for the one-layer geobelt reinforced cushion, while for the two-layer geobelt reinforced cushion, the ratios were ranging from 2.46 to 2.63.
Figure 17 shows the relationship between the depth of the rst geobelt layer beneath the footing (U) and SDR. Figure 18 shows the relationship between the linear density of reinforcement (LDR) and SDR.As can be seen, U and LDR had little impact on stress di usion of the cushion.e possible reason is that the depth of the cushion adopted in 8 Advances in Materials Science and Engineering this study was small.As can be seen in Table 6, at the proportional limit state, the SDR decreased from 1.50 to 1.46 as U increased from 50 mm to 200 mm, while at the stable state, the SDR increased from 1.76 to 1.84 as U increased from 50 mm to 200 mm.It should be noted that the relationship between SDR and U is consistent with that between BCR and U. is also indicates that the geobelt only takes its e ect e ciently after certain amount of deformation of cushion happens.e SDR was found to increase from 1.47 to 1.48 at the proportional limit state and increase from 1.77 to 1.84 at the stable state as the LDR increased from 25% to 50%.erefore, the LDR showed unnoticeable e ect on the stress di usion of the geobelt reinforced cushion.Figures 19(a) and 19(b) show the e ect of vertical spacing between two geobelt layers (H) on the stress di usion angle.For di erent types of geobelt, the values of SDR showed a similar trend, which is the stress di usion angle increased slightly with the increase of H.At the proportional limit state, the SDRs of the TG-reinforced cushion were 1.57 and 1.60, while for the CPE-reinforced cushion, the values were 1.51 and 1.52, which is slightly smaller than that of the TGreinforced cushion.At the stable state, the SDRs of the TG-reinforced cushion were 2.54 and 2.63, while for the CPEreinforced cushion, the values were 2.46 and 2.60, which was also slightly smaller than that of the TG-reinforced cushion.It can be concluded that the reinforcing e ect of TG geobelt was slightly better than that of CPE geobelt.e results are also consistent with BCR results discussed in Section 3.1.

Engineering Applications
In this section, a practical project included in Technical code for ground treatment of buildings [31] was presented to demonstrate the e ect of geobelt.
e site of the project is located on a saturated silt soil with a bearing capacity of 80 kPa.e footing has a length and width of 60.8 m and 14.9 m, respectively.e thickness of the footing is 2.73 m.An average pressure beneath the footing is 130 kPa.Two possible treatment solutions are (1) unreinforced gravel cushion with a density of 19.5 kN/m 3 , and the thickness of the cushion is taken as 3.73 m by calculation and (2) two-layer TG geobelt-reinforced thin gravel cushion.
e density of the cushion was 19.5 kN/m 3 .e width (B) was 17 m and thickness (Z) was 1.5 m with a Z/B ratio of 0.088 (<0.2). e project adopted the second treatment solution by using the geobelt-reinforced cushion.e rst layer geobelt beneath the cushion was 0.6 m, and the vertical spacing between two layers was 0.4 m. e linear density reinforcement is 17%.According to the test results, the stress di usion angle takes 35 °considering a certain safety factor.After a comparative analysis, the reinforced cushion used in the second solution reduced the cushion thickness by 60% compared to the unreinforced cushion.e deformation after engineering settlement was 3.9 mm which meets the requirements.
According to this project, the stress di usion angle of the reinforced gravel cushion was larger than that of the unreinforced one, which satis es the bearing capacity and deformation requirements.In addition, the thickness of the cushion and construction cost can be reduced by using geobelt reinforcement.

Conclusions
By conducting the in situ square footing loading test on the geobelt-reinforced thin gravel cushion, this study investigated the reinforcing e ect of geobelt on bearing capacity of the foundation and stress di usion of cushions.e following conclusions can be made:   2) e earth pressure distribution on bottom of reinforced cushions was in a saddle shape, which was smoother than that of unreinforced cushions.When loaded, geobelt can restrict the movement of soils due to interfacial frictions between it and crushed gravels; therefore, the stress distribution of reinforced cushions becomes less varied compared to that of unreinforced cushions.(3) e deformation and stress distribution of underlying soil layers under reinforced cushions were     Advances in Materials Science and Engineering much di erent from that under unreinforced cushions.Especially, the e ect of geobelt on stress di usion was pronounced.e stress di usion angles of reinforced cushions were all larger than 20 °.
(4) e reinforcing e ect of two-layer geobelt reinforcement is obviously better than that of one-layer geobelt reinforcement.At the stable state, the stress di usion angles of one-layer reinforced cushions were increased by 1.76-1.84times while two-layer reinforcement was improved by 2.46-2.63times, compared to unreinforced cushions.
(5) At the stable state, for the one-layer geobelt reinforced cushion, the stress di usion and bearing capacity can be improved by increase of distance between the rst geobelt and footing.A higher linear density of reinforcement can get a better bearing capacity and stress di usion.(6) At the stable state, for the two-layer geobelt reinforced cushion, the greater the vertical spacing between two layers, the better the stress di usion e ect and the larger the bearing capacity of the reinforced cushion.And the performance of the reinforced cushion with TG geobelt was better than that of the reinforced cushion with CPE geobelt.

Figure 1 :Figure 2 :
Figure 1: Setup of the loading test.

Figure 15 :
Figure 15: Stress di usion of the cushion.

Table 1 :
Physical and mechanical parameters of silt soil.

Table 3 :
Engineering properties of the geobelts.

Table 4 :
Physical parameters of gravel.

Table 5 :
e summary of test results.

Table 6 :
e summary of stress di usion angles.