Mechanical mechanism of soft soil foundation improved by geosynthetic-reinforced pile-supported technology based on field experiment

Geosynthetic-reinforced pile-supported (GRPS) technology is an effective method for improving soft soil foundations. Conducting field tests is crucial for accurately assessing the mechanical behavior of GRPS technology. Based on the fourth expansion project at Pudong Airport, field tests were conducted to investigate the effects of GRPS technology on soft-foundation treatment. Specifically, the tests examined the changes in the pile and soil settlement, pore water pressure, and pile and soil stress. The results showed that the settlement of the GRPS foundation mainly occurred during the loading period, accounting for approximately 65% of the total foundation settlement. Increasing the pile modulus and reducing the pile spacing decreased the total foundation settlement. The filling load was mainly transferred to the piles through the soil arching effect, resulting in a large excess pore water pressure in the soil layer near the pile tip. In addition, when the filling height was low, the load transfer relied predominantly on the tensile membrane effect of the geogrid, which was more pronounced near the piles. However, as the filling height increased, the soil arching effect surpassed the tensile membrane effect and played a key role in load transfer.


Introduction
With the rapid development of China's transport industry, the construction of highways, high-speed railways, and airport runways has inevitably been subject to challenges posed by problematic deep soft foundations.These soft foundations, which are unsuitable for supporting roads and runways, usually require treatment because of their insufficient bearing capacity, settlement, differential settlement, and excessive lateral deformation [1,2].Conventional foundation treatment methods, such as strong tamping, filling replacement, and precompression consolidation, have significant limitations in terms of treatment depth and construction period [3].In this regard, the geosynthetic-reinforced pilesupported (GRPS) foundation has proven to be an effective alternative as it enhances the bearing capacity and reduces post-construction settlement and differential settlement by facilitating stress coordination between the piles and the foundation soil.Moreover, GRPS technology offers a short construction period, making it particularly suitable for reinforcing the deep soft foundations of airports.
In recent years, extensive research has been conducted by scholars, both domestically and internationally, on the GRPS technology from various perspectives, including theoretical analysis, 1336 (2024) 012007 IOP Publishing doi:10.1088/1755-1315/1336/1/012007 2 model testing, and numerical simulations.In theoretical research, some scholars have proposed vertical shear surface models, semicircular or spherical soil arching models, unequal radius spherical shell-shaped multiple arching models, and concentric circle arching models based on observed experimental phenomena and certain assumptions.Theoretical design methods have also been developed for these models [4][5][6][7][8].Zhao et al. [9] improved the design calculation theory for GRPS embankments by considering the relative interaction between the pile and the soil in a soil column model.In model testing, Xu et al. [10] analyzed the effects of fill material properties and clear pile spacing on the soil arching effect through laboratory-scale model tests.Rui et al. [11] conducted multitrapdoor tests based on two-dimensional steel rod with similar soils to investigate the deformation patterns and soil arching evolution of pile-supported embankments under different parameter combinations.Zheng et al. [12] investigated the mechanical behavior of GRPS embankments in medium-and low-compressibility soils through field tests.Their findings indicated that the tensile membrane effect on the shoulder was greater than the soil-arching effect.Yang et al. [13] compared five common GRPS embankment design methods, based on field tests and monitoring data.Their study revealed that the pile-soil sharing ratios calculated using different theoretical design methods were different.Using numerical analysis, Fei et al. [14] and Zhuang et al. [15] revealed the mechanism of the soil arching effect and the tensile membrane effect of the GRPS embankment based on ABAQUS numerical simulation and found that different theoretical design methods were applicable to the change in embankment filling height.Lai et al. [16] and Bao et al. [17] used the PFC discrete element method to explain the mechanical transfer mechanism of GRPS embankments from the perspective of microscopic force chain changes.Numerous studies have been conducted on GRPS embankments; however, consistent conclusions have yet to be reached.Field tests, provide a relatively realistic reflection of the mechanical mechanism of GRPS technology; however, related research in this area remains relatively scarce.
Therefore, field tests were conducted to investigate the mechanical mechanism of the GRPS technology in treating soft foundations, specifically focusing on the fourth expansion project at Pudong Airport.This study mainly compared the differences between triaxial mixing piles and PHC piles.The changes in the pile and soil settlement, pore water pressure in the foundation, and stress distribution of the pile and soil between the piles were analyzed.These results are helpful in revealing the mechanical mechanism and accumulating experience for the application of GRPS technology in the treatment of deep soft foundations in airports.

Project overview
A foundation treatment test was conducted at a site located between the old and new embankments at Pudong Airport.The foundation soil at a depth of 35.45 m consists of Quaternary, Late Pleistocene, and Holocene sediments.According to a geological investigation report, the foundation soil primarily consists of clay, powdery, and sandy soils with a layered distribution pattern, as depicted in figure 1.The stable water level depth ranges from 0.20 m to 1.45 m below the ground surface, and the physical and mechanical parameters of each soil layer were presented in table 1.
3    After the strength of the pile cap met the required standard, a cushion layer consisting of graded broken stones with a thickness of 50 cm was placed on top of the pile cap.Two layers of TGSG5050 type geogrid were installed on the surface of the pile cap and in the middle of the cushion layer.The horizontal lap of the geogrid was not less than 20 cm and the vertical lap was not less than 50 cm.The filling height was 3 m with a unit weight of 18 kN/m 3 , as illustrated in figure 4.    2 and figure 3, respectively.The settlement plates marked DB-1, DB-2, DB-5, and DB-6 were placed at the top of the piles to monitor the settlement of the pile, whereas DB-3, DB-4, DB-7, and DB-8 were placed on the soil surface between the piles to monitor the settlement of the soil.Two pore water pressure holes, KX-1 and KX-2, were placed inside and outside the test area, respectively.Within each hole, six water pressure meters were installed at 5-m intervals, starting at a depth of 2 m.Soil pressure boxes were positioned at the plane of the pile cap, first layer of the geogrid, and second layer of the geogrid.The soil pressure boxes on the pile were marked as ZY-1 to ZY-3, whereas those on the soil between the piles were marked as TY-1 to TY-3, as shown in figure 2(a) and figure 3(a).The triaxial mixing pile had a construction period of 90 days and preloading period of 370 days, whereas the PHC pile had a construction period of 70 days and preloading period of 340 days.

Variation of surface settlement
Figure 5 illustrates the variation in the surface settlement at the top of the pile and the soil between the piles as the filling height changed in both the triaxial mixing pile and PHC pile areas.The figures reveal a close correlation between the settlement and loading processes.Initially, in the early loading stage, the settlement of both the pile and soil was relatively slow owing to the limited filling height, with settlement rates remaining below 1 mm/d.As the filling height increased, the settlements of both the pile and soil gradually increased.Notably, the soil settled faster than the pile.During the preloading stage, the settlement variation of the pile and soil slowed and tended to stabilize as the foundation soil consolidated.The settlement of both the pile and the soil between the piles primarily occurs during the loading period.At the end of the loading phase, the settlement of the pile and soil between the piles in the triaxial mixing pile area reached 80 mm and 120 mm, respectively, accounting for approximately 61% of the total foundation settlement.In the PHC pile area, the settlement of the pile and soil between the piles reached 70 mm and 130 mm, respectively, accounting for approximately 68% of the total foundation settlement.Figure 5 also shows that as the filling height increases, the settlement of soil between the piles surpasses that of the pile, suggesting the presence of an initial soil arching effect during the filling process.This is because the differential settlement between the piles and soil causes stress deflection in the filling materials and the formation of a soil arching effect.However, the differential settlement between the piles and soil was lower during the filling process, and only the initial soil-arching effect remained.The foundation settlement continued after the end of the loading because the piles did not penetrate the soft soil layer.The compression of the soft soil under the pile bottom and the consolidation of the foundation soil resulted in the synchronous settlement of both the pile and soil between the piles during the preloading period.It is worth noting that the PHC pile has a larger modulus and length compared to the triaxial mixing pile.Consequently, the total settlement and post-construction settlement in the PHC pile area were smaller than those in the triaxial mixing pile area.Additionally, for the same pile type, figure 5(a) shows that the settlement of soil between piles in S1 was larger than that in S2, whereas figure 5(b) shows that the settlement of soil between piles in Y2 was larger than that in Y1.Therefore, by combining table 2 and table 3, it can be concluded that increasing the pile modulus and reducing the pile spacing significantly contribute to the reduction of the total settlement in the foundation.In GRPS foundation structures, the occurrence of the soil arching effect during filling can be attributed to the differential settlement between the piles and the soil between the piles.The variation in the pile-soil differential settlement over time in the triaxial mixing and PHC pile areas is shown in figure 6.As shown in figure 6, the pile-soil differential settlement was minimal when the filling height was relatively small.However, as the filling height increased, the pile-soil differential settlement in each region rapidly increased.Once the filling height exceeds 1.4 times the clear pile spacing, the rate of the differential settlement increment notably slowed, and the differential settlement gradually reached its peak value before gradually decreasing.After the completion of loading, there was a slight increase in the pile-soil differential settlement; however, the magnitude of this change was relatively small.The variation in the pile-soil differential settlement over time reflects the evolution of the soilarching effect.As the pile-soil differential settlement increases, the initial soil arching effect gradually emerges in the filling, facilitating the transfer of the filling load to the pile.Consequently, the settlement of the pile increases, and the rate of the pile-soil differential settlement increment decreases.Once the complete soil arching was established, the increased load was fully borne by the pile.During this stage, the settlement of the pile surpassed that of the soil between the piles, leading to a decrease in the pile-soil differential settlement.The increase in the pile-soil differential settlement during the preloading period indicated that the soil between the piles underwent consolidation settlement.Additionally, figure 6 shows that the rate of increase in the pile-soil differential settlement is higher in the PHC pile area than in the triaxial mixing pile area.When the differential settlements were equal, the filling height in the PHC pile area was smaller.Similarly, when the filling height was the same, the peak value of the differential settlement in the PHC pile area was larger.The main reason for this is that the length and modulus of the PHC piles were significantly larger than those of the triaxial mixing piles, and a complete soil arching effect was more likely to occur because of the smaller clear pile spacing in the PHC pile area.In contrast, the peak values of the differential settlement in areas S1, S2, and Y2 were almost the same, whereas this value was significantly larger in Y1 area, which indicates that the effect of changing the pile spacing on the soil arching effect is more pronounced than that of increasing the pile modulus.In it ia l a rc h in g Variation curves of pile-soil differential settlement.

Variation of pore water pressure
The variations in the pore water pressure over time, both within and outside the test areas of the triaxial mixing pile and PHC pile, are depicted in figure 7 and figure 8, respectively.During the loading period, the pore water pressure within the foundation gradually increased with the filling height.Once loading was completed, the increase in pore water pressure reached its peak value.
During the preloading period, the progressive dissipation of pore water leads to a reduction in the pore water pressure, thereby facilitating consolidation and settlement.Figure 7(a) indicates that the maximum increase in the pore water pressure in the triaxial mixing pile area was 22.5 kPa, occurring at a depth of 17-22 m in the soil layer.Figure 8(a) shows that the maximum increase in the pore water pressure in the PHC pile area was 44 kPa, occurring at a depth of 27 m.This can be attributed to the disparity in pile length, with the triaxial mixing piles measuring 21 m and the PHC piles 26 m.The increase in the pore water pressure was caused by an increase in the load.This phenomenon indicates that the load from the filling was predominantly transmitted to the piles through the soil arching effect, causing a substantial excess pore water pressure in the soil layer near the pile tip.Additionally, the increase in pore water pressure in the PHC pile area surpassed that in the triaxial mixing pile area, primarily because of the smaller pile spacing in the PHC pile area, thereby intensifying the soil arching effect.This result is in accordance with that of the settlement.A notable discrepancy in the location of the maximum increase in the pore water pressure can be observed by comparing the pore water pressure both within and outside the test area for the same type of pile.In the test area, the piles transferred the load and generated a significant excess pore water pressure within the deep soil layer at depths of 17 m, 22 m, and 27 m.Conversely, the increase in the pore water pressure within the shallow soil layer was relatively minor, as shown in figure 7(a) and figure 8(a).However, outside the test area, the results exhibited the opposite trend.The pore water pressure increased more significantly in the shallow soil layer above a depth of 7 m, while the change in the pore water pressure in the deeper soil layer below 17 m was less pronounced, as illustrated in figure 7(b) and figure 8(b).The reason for this phenomenon is that the change in the pore water pressure outside the test area was mainly affected by the filling load in the test area, and this effect on the shallow soil layer was relatively obvious.In addition, in the triaxial mixing pile area, the excess pore water pressure in each soil layer tended to dissipate completely by the end of preloading.However, in the PHC pile area, a residual excess pore water pressure of 10 kPa remains in the deep soil layer below 17 m.This can be attributed to the fact that the foundation soil below the 17-m depth consists of clay and silty clay, characterized by a low coefficient of permeability and a lengthy drainage path, thereby resulting in a slower rate of pore water pressure dissipation.

Variation of pile and soil pressure
The variation in soil pressure on the pile and between the piles over time in both the triaxial mixing pile and PHC pile areas is shown in figure 9.The figure also presents the variation in filling height over time for analytical convenience.As shown in the figure, as the filling height increased, the soil pressure on the pile and the soil between the piles increased to varying degrees.Initially, at a low filling height, the rates of increase in soil pressure on the pile and the soil between the piles were similar, resulting in minimal differences in the soil pressure values.However, as the filling height gradually increased, the rate of increase in the soil pressure on the pile accelerated, leading to a gradual divergence in the soil pressure values between the piles.At the maximum filling height, the soil pressure on the pile and that between the piles reached their peak values.During the preloading period, the soil pressure on the soil between the piles decreased slightly because of foundation soil consolidation, whereas the soil pressure on the piles fluctuated around the peak value.Comparing figure 9(a) and figure 9(b), the variation in soil pressure on the pile and soil between piles with the filling height was consistent in both the triaxial mixing pile and the PHC pile areas.However, the values of soil pressure on the piles and soil between the piles after the completion of filling were not equal.In the triaxial mixing pile area, the soil pressure on the pile at the plane of pile cap ZY-3 was 103 kPa, approximately 1.9 times the average soil pressure of the filling.The soil pressure on the soil between the piles at TY-3 is 28 kPa, approximately 0.5 times the average soil pressure of the filling.In the PHC pile area, the soil pressure on the pile at the plane of pile cap ZY-3 was 120 kPa, approximately 2.2 times the average soil pressure of the filling.The soil pressure on the soil between the piles at TY-3 is 38 kPa, around 0.7 times of the average soil pressure of the filling.The soil pressure on the pile in the PHC pile area was greater than that in the triaxial mixing pile area, whereas the soil pressure on the soil between the piles in the PHC pile area was lower than that in the triaxial mixing pile area.Therefore, the soil arching effect was more pronounced in the PHC pile area, which agrees with the settlement analysis results.
In addition, the data in figure 9 indicate that the soil pressure on the pile at the plane of the pile cap was the highest, followed by the first layer of the geogrid, and the soil pressure on the pile at the second layer of the geogrid was the lowest.Conversely, the soil pressure exhibited the opposite trend.The soil pressure on the soil between the piles in the second layer of the geogrid was the highest, and the value at the plane of the pile cap was the lowest.This observation suggests that, during the downward transfer of the filling load, the soil pressure that should be sustained by the soil is further transferred to the pile owing to the tensile membrane effect of the geogrids in the cushion layer.After the completion of loading, the soil pressure on the pile in the first layer of the geogrid increased by approximately 5-7 kPa compared with that in the second layer of the geogrid, and the soil pressure on the soil decreased by approximately 2-5 kPa.The soil pressure on the pile at the plane of the pile cap increased by approximately 12-15 kPa compared with the first layer of the geogrid, and the soil pressure on the soil decreased by approximately 5-8 kPa.Evidently, more of the load was transferred from the first layer of the geogrid to the plane of the pile cap.This indicated that the tensile membrane effect of the geogrid near the pile cap was more pronounced.

Variation of load sharing
Based on the above analysis, it is evident that the downward transfer of the filling load in the GRPS foundation structure is influenced by both the soil arching effect in the filling and the tensile membrane effect of the geogrid in the cushion layer.To further investigate these effects, it was assumed that the difference between the soil pressure on the pile in the second layer of the geogrid and the average soil pressure of the filling represented the load shared by the soil arching effect.Similarly, the difference between the soil pressures on the pile in the first and second layers of the geogrid represents the load shared by the tensile membrane effect of the second layer of the geogrid.Moreover, the difference between the soil pressure on the pile at the plane of the pile cap and the first layer of the geogrid represents the load shared by the tensile membrane effect of the first layer of the geogrid.Consequently, the load sharing of the soil arching and tensile membrane effects of the two layers of the geogrid in the triaxial mixing pile and PHC pile areas were calculated, as shown in figure 10.As observed in figure 10, as the filling height increases, the loads shared by the soil arching and tensile membrane effects gradually increase in the triaxial mixing pile area.In contrast, the loads that shared these effects in the PHC pile area initially increased and then stabilized.During the early stage of loading, the filling height was low, and there was almost no soil arching effect on the filling.The load shared by the tensile membrane effect of the two geogrid layers was small.However, as the filling height increased, the load shared by the soil-arching effect increased rapidly.In contrast, the increase in the load shared by the tensile membrane effect was relatively small, particularly in the PHC pile area.Once a certain filling height was reached, the capacity for load sharing owing to the tensile membrane effect became relatively stable.These findings reveal that load transfer is primarily governed by the tensile membrane effect when the filling height is low in the GRPS foundation structure.However, as the filling height increases, the soil arching effect gradually surpasses the tensile membrane effect and becomes the dominant factor in the load transfer process.At the completion of loading, the loads shared by the soil arching effect, tensile membrane effect of the first layer of the geogrid, and tensile membrane effect of the second layer of the geogrid in the triaxial mixing pile area were 14.7 kPa, 12.8 kPa, and 6.2 kPa, respectively.In the PHC pile area, the loads shared by the soil arching effect, the tensile membrane effect of the first layer of geogrid, and the tensile membrane effect of the second layer of geogrid are 53.3 kPa, 18.5 kPa, and 12.3 kPa, respectively.These results indicate that the soil arching effect possesses the greatest load-sharing capacity in both GRPS foundation structures, followed by the tensile membrane effect of the first layer of the geogrid.The tensile membrane effect of the second layer of the geogrid exhibited the smallest load-sharing capacity.The load-sharing capacities of the soil arching and tensile membrane effects of the first layer of the geogrid in the triaxial mixing pile area were close but significantly smaller than the load-sharing capacity of the soil arching effect in the PHC pile area.This difference could be attributed to the larger modulus, longer length, and smaller clear pile spacing in the PHC pile area, which resulted in a greater soil-arching effect.

Variation of pile-soil stress ratio
The pile-soil stress ratio is a crucial indicator that reflects the load-transfer mechanism of the GRPS foundation structure.It is defined as the ratio of the soil pressure on the pile to the soil pressure on the soil between the piles.The variation in the pile-soil stress ratio with the filling height in both the triaxial mixing pile and PHC pile areas is shown in figure 11.In the triaxial mixing pile area, the pilesoil stress ratios at the plane of the pile cap, as well as at the first and second layers of the geogrids, increased linearly with the filling height.The pile-soil stress ratios reached 3.1, 1.9, and 1.6, respectively, at the end of loading.In the PHC pile area, the pile-soil stress ratios initially increased with the filling height and gradually stabilized at a certain level.The pile-soil stress ratios reached 5.7, 3.5, and 2.9, respectively, at the end of loading.Notably, the pile-soil stress ratio at the plane of the pile cap was approximately 1.6 and 1.9 times that of the stress ratios at the first and second geogrids, respectively.This difference became more significant as the filling height increased.The difference in the pile-soil stress ratios between the first and second geogrids was small, suggesting that the load was primarily transferred to the pile through the soil arching effect, with limited impact from the tensile membrane effect.In addition, the pile-soil stress ratios at various locations in the PHC pile area were higher than those in the triaxial mixing pile area.This is because of the greater pile-soil differential settlement in the PHC pile area, which leads to a stronger soil arching effect and tensile membrane effect, resulting in a higher transfer of load from the soil between the piles to the piles.

Conclusions
In this study, field tests were conducted to investigate the effectiveness of GRPS technology in treating soft foundations.The variations in the pile and soil settlement, pore water pressure, and pile and soil stresses in the triaxial mixing pile and PHC pile areas were investigated.The main conclusions are as follows.
(1) The settlement of the soft foundation treated with GRPS technology occurred primarily during the loading period, accounting for approximately 65% of the total foundation settlement.As the filling height increased, the settlement of the soil between the piles became larger than that of the piles.An initial soil arching effect was generated during the filling process, and a complete soil arch was gradually formed.During the preloading period, the soil arch remains relatively stable.Increasing the pile modulus and reducing the pile spacing can reduce the total foundation settlement, and changing the pile spacing has a more significant effect on the soil arching.
(2) In both the triaxial mixing pile and PHC pile areas, the filling load was primarily transferred to the piles through the soil arching effect, leading to a large excess pore water pressure in the soil layer near the pile tip.The variation in the pore water pressure outside the test area was mainly affected by the filling load within the test area, and the effect on the shallow soil layer was more obvious.
(3) When the filling height was low, the increase in soil pressure on the pile and the soil between the piles occurred at a similar rate, resulting in a small difference in the soil pressure values.As the filling height increased, the difference between the pile and soil pressures gradually increased.During the downward transfer of the filling load, the soil pressure that should be borne by the soil between the piles is transferred to the piles because of the soil arching effect in the filling and the tensile membrane effect of the geogrid in the cushion layer.The tensile membrane effect of the geogrid was more pronounced close to that of the piles.
(4) In a GRPS foundation structure, the load transfer primarily occurs through the tensile membrane effect of the geogrid when the filling height is low.However, as the filling height increases, the soil arching effect surpasses the tensile membrane effect and plays a key role in load transfer.At the end of the loading, the pile-soil stress ratio at the plane of the pile cap was approximately 1.6 times and 1.9 times of the pile-soil stress ratio at the first and second layers of the geogrids, respectively, indicating that the load-sharing capacity of the soil arching effect was the highest, followed by the tensile membrane effect of the first layer of the geogrid, with the second layer of the geogrid exhibiting the least influence.

Figure 1 .
Figure 1.Soil profiles at the test site.

Figure 2 .Figure 3 .
Figure 2. Test area of triaxial mixing pile: (a) cross section of A-A; (b) plan view.
Settlement of pile in S1 Settlement of pile in S2 Settlement of soil between piles in S1 Settlement of soil between piles in S2 Settlement of pile in Y1 Settlement of pile in Y2 Settlement of soil between piles in Y1 Settlement of soil between piles in Y2

Figure 7 .
Figure 7. Variation of pore water pressure in the triaxial mixing piles area (a) within the test area; (b) outside the test area.

Figure 8 .
Figure 8. Variation of pore water pressure in PHC piles area (a) within the test area; (b) outside the test area.

Figure 9 .
Figure 9. Variation of pile and soil pressure (a) triaxial mixing piles area; (b) PHC piles area.
Tensioned membrane effect of the first geogrid Tensioned membrane effect of the second geogrid (a) and (b).
The plane of pile top The plane of the first geogrid The plane of the second geogrid(a)   and (b).

Table 1 .
Physical and mechanical parameters of foundation soils.

Table 2 .
Parameters of the triaxial mixing pile.

Table 3 .
Parameters of the PHC pile.