Abstract
Although pile-net composite foundation has been widely used in the construction of high-speed railway, load transfer characteristics of its structural system during the filling of roadbed are seldom studied, so field studies are conducted in the Gan-Long (Ganzhou-Longyan) high-speed railway DK86 + 998.0–DK87 + 191.2 test section, aiming to explore the bearing characteristics of a short pile-net composite foundation over low to medium compressible soil. The study encompasses a field measurement of Earth pressure and pore water pressure which lasted almost two years, and a systematic analysis of variation law of soil pressure between piles and their tops and that of pore pressure in reinforcement zone and retaining layer in medium and low compressibility soil area. The results suggest that the soil in the test section bears the characteristic of a low water content, low porosity ratio, and high liquid limit, with compression factor being approximately 0.25, and test values of the natural soil pressure and the lateral distribution are close to the theoretical values. Soil pressure at the top of the piles is higher than that between the piles by approximately 2.4 times because of the differences in pile-net rigidity. Besides, the soil pressure of the pile-net composite foundation follows the pattern of “jagged” in the transverse direction of the roadbed. The filling load of the composite foundation soil arch is 80 kPa and the composite foundation reaction coefficient is 6.4 kN/m3 when the filling is stable. The pile-soil stress ratio at the shoulder and the center of the line is 3.2 and 2.8, respectively. The change of the hole pressure at the pile end point is larger than that at the reinforcement area, and the side verification short pile can effectively transmit the additional stress of the upper part to the pile end soil layer.
1. Introduction
Because of the climate differences between the north and south regions in China, there are many difficulties in the design and construction of high-speed railway foundations in different districts [1–4], so composites with better performance are designed in the construction process of high-speed railway. Among these composite foundations, the pile-net composite foundations are widely used because the characteristics of horizontal reinforcement composite foundation and vertical reinforcement composite foundation could be combined [5–7]. The pile-net composite foundation consists of six main parts: upper embankment filling, grille mesh, gravel cushion, pile-soil composite foundation reinforcement area, soft soil layer, and pile bearing layer [8, 9], and the piles, net, and soil work together to bear the upper load. In order to solve different complex geological conditions during the construction of high-speed railway in China, it is of great significance to study the bearing characteristics of pile-net composite foundation for the engineering practice.
At present, many researchers have studied the characteristics of pile-net composite foundations using experimental methods and numerical modelling [10, 11]. Bi et al. [12] carried out two geotechnical centrifuge model tests to investigate the reinforcement effect of inclined piles for the existing pile-net composite foundation on soft soil overlying sloping base. Jiang et al. [13] improved of the current pile-supported embankment calculation model from soil arching effect, geotextile deformation, and pile-soil load transfer. Huang et al. [14] adopted the equivalent stiffness ideal elastic-plastic model for the transfer function of relative displacement surface on the basis of Terzaghi model. Chen et al. [15] analyzed the mechanical properties and design-calculation methods of pile-net composite foundation. In China, the research on the pile-net composite foundations has also been conducted based on the various practical projects. Based on the Kunshan test section, Jing-Hu high-speed railway, Hu-Ning intercity railway, and other projects, field tests, theoretical analyses, and numerical simulations were carried out [16–18]. Additional investigations, including assessment of the reinforcing effect, the time and space characteristics of sedimentation, the reaction characteristics, and usage of the new concept of “vulnerability assessment” were also performed to evaluate the working state of the foundation during its operational life [19, 20].
Meanwhile, considering that settlement control has become one of the most important technologies in the design of high-speed railway, it is necessary to study the engineering mechanical properties of different types of soils. Su et al. [21] analyzed the hydrological and mechanical properties of diatomaceous Earth on the basis of a field survey and laboratory in the new Hang Shaotai high-speed railway project. Fu and Yuan [22] investigated the dynamic response of composite foundation in soft soil for a ballastless high-speed railway. Shao et al. [23] carried out four immersion tests of sand wells with different depths in a loess site on the Baoji-Lanzhou high-speed railway. Pan et al. [24] presented an experimental study on a natural UK stiff clay which existed in London. Soils with low to moderate compressibility are widely distributed in China, and most of high-speed railway foundations are constructed on these soils. However, the reinforcement technology of medium and low compressibility soil foundation is not mature at present [25–29], and bearing characteristics of a short pile-net composite foundation over soil with low to moderate compressibility continues to be an open problem.
Therefore, a field test is conducted in the Gan-Long (Ganzhou-Longyan) test section to investigate the bearing characteristics of a moderately compacted short pile-net composite foundation, which is more accurate when compared with laboratory test and numerical simulation. Firstly, the railway project is introduced briefly and geological conditions in the test section are described. Then, field test method is designed, which contains the usage of different equipment and the layout of the test elements. Finally, results of the field test, including the development of Earth pressure, characteristics of Earth pressure distribution along the base, soil arch effect in foundation filling, and so on are discussed. The findings of this study can help for better understanding of the bearing characteristics of the moderately compacted short pile-net composite foundation and provide guidance for similar projects.
2. Field Test Methods
2.1. Overview of the Project
The length of the Gan-Long (Ganzhou-Longyan) railway is 290.1 km. The construction of this railway began in 2010 and opened to traffic in 2016. The designed speed is more than 250 km/h. The test site is near Xiaomi town in Yudu County, Ganzhou City. The test site location is DK 86t + 998.0–DK 87 + 191.2 with a length of 192.88 m (including a short chain of 0.32 m). This section has a gentle slope in front of a mountain and the terrain is gentle with mostly copse and regolith and some houses. The surface is a grayish yellow, hard plastic conglomeratic clay with a thickness of 18–32 m underlain by C2h limestone with well-developed caves.
Surface water in the test section is mainly derived from runoff water, and the groundwater is mainly derived from the carbonate rock, karst water, and pore phreatic water. The depth of the pore phreatic water is 1.5 m–12.7 m, which is easily influenced by seasonal changes. The karst-fissure water, whose water storage capacity is influenced by the development of karst fissures, is unevenly distributed. The quaternary strata directly overlie carbonate rocks in this area, so there is a close hydraulic connection between the surface water and the groundwater. The pore phreatic water and the carbonate karst-fissure water are continuous and both the surface water and the groundwater are nonerosive.
According to the comprehensive analysis of drill and cone penetration tests, plate loading tests, and some indoor experiments, the stratigraphic distribution from top to bottom in the test section is as follows:(1)Silty clay (Q4al + pl): Yellowish dark brown-brownish yellow, hard plastic, mostly cohesive soil equally distributed. The cut surface is relatively smooth with 2–5% gravel. The grain sizes range from 2 to 10 mm, most of which is sand and small rounded gravel; the thickness is 0.5–15 m; and the bearing capacity is 160 kPa.(2)Silty clay (Qel + dl): Brownish yellow, hard plastic, mostly cohesive soil equally distributed with some Fe-Mn nodules. Its cut surface is rough with 5–30% gravel. The grain sizes range mostly from 2 to 30 mm with some as large as 60 mm. Most are weathering fragments with angular shapes and fragmental structure. The thickness is 8–25 m, and the bearing capacity is 180 kPa.(3)Limestone (C2h): Steel-gray with weak weathering, karst development, and bearing capacity of 800 kPa.
Combining the results of in situ experiments and indoor geotechnical tests, all the physical mechanical parameters of the soils in the test section are summarized. Details are listed in Tables 1–3. From Tables 1–3, it can be found that the physical properties in all the layers in the foundation soil have high homogeneity. The variability coefficient of all the physical property indices is relatively small and does not change much with depth. The natural moisture content is approximately 24%, the void ratio is approximately 0.7, and the liquid limit is approximately 40%. Meanwhile, the soil has a low moisture content, low void ratio, and high liquid limit. The compressibility coefficient is relatively stable, and all of them are approximately 0.25, which denotes medium compressible soil.
Considering the soil characteristics in the test section, preloading (the preloading pillar is 2.0 m in height) is used in DK 87 + 0.700–168.0. The pile-net composite foundation is square with 1.6 m pile spacing, 0.5 m pile diameter, and 6 m pile length. Polypropylene is placed on top of the pile subgrade (0.4 m thick crushed subgrade sand + 0.2 m medium-coarse sand) to stretch the TGDG (TGDG 100 kN/m), with an elongation less than 10%. The width of roadbed is 13.2 m, and the side slope is 1 : 1.5.
2.2. Field Test Methods
The testing elements were put in section DK 87 + 155 and DK 87 + 185, where foundation settlement, layered settlement, foundation lateral (horizontal) displacement, pore water pressure, and soil pressure were tested. The DL-502 electronic balance level was used to test settlement. A YH6406 reader was used to test stratum settlement, and JM2X-5503AT pore water pressure gauges were used to test the change of groundwater pore pressure. A soil pressure box, JMZX-5020AT, was placed between the pile top and the soil among piles. The layout of the test elements is shown in Figure 1.

The test elements were installed on November 6, 2012, and the tests began on January 26, 2013, and ended on January 6, 2015. The testing frequency during foundation construction was once per day and 2-3 times per day during compaction. When the foundation construction was completed, the testing frequency during the settlement period was twice per week in the first 3 months and once per month after the first 3 months.
3. Results and Discussion
3.1. Development of Earth Pressure
Figure 2 shows the curve of the Earth pressure-time-load based on the natural foundation and pile-net composite foundation test, and it could be found that the Earth pressure and filling load are consistent with the change of time. In the beginning period, the load and the Earth pressure are both small and the change of the curve is mainly influenced by construction machines, which causes the difference between load and Earth pressure. In the second stage, the difference between Earth pressure and filling load gradually decreases, which is caused by the increase of embankment filling height and the decrease of Earth pressure under the influence of construction mechanical load. Similar to the relationship between filling load and Earth pressure in the second stage, the Earth pressure of Sp-2-5, Sp-2-7, and Sp-2-8 is 103 kPa, 154 kPa, and 120 kPa, respectively, in the third period, which is 0.8, 1.1, and 0.9 times the filling load. When the filling load is stable, the Earth pressure value is decreased with increasing distance of the measuring point from the subgrade center, because the load of embankment is in the form of trapezoidal load distributed in the foundation, and the foundation at the shoulder and slope toe is subjected to the triangular load in the trapezoidal load, so the test value is less than the fill load. The foundation is formed into a settlement basin under the filling load, and the deformation of roadbed is extruded at the center to form stress concentration, so difference between Sp-7 and Sp-8 at the same measuring points inside the shoulder could be observed.

(a)

(b)
The similarity between the composite foundation and natural foundation is that the soil pressure is consistent with the change trend of filling load curve, and the initial test of filling is susceptible to the influence of construction machinery load. There is a positive correlation between the soil pressure in the composite foundation testing point and at the center distance from its foundation. The Earth pressure in the center of the testing foundation is the largest, followed by the central line and the shoulder. The error between the theoretical value and the measured value in the central testing period is 15%. Because there are differences in pile-net rigidity, and the pile with larger stiffness is prone to stress concentration during embankment filling, the soil pressure at the top of the piles is higher than that between the piles by approximately 2.4 times. The error between the theoretical value and the measured value in soil pressure in the central foundation is 6.67%.
3.2. Characteristics of Earth Pressure Distribution along the Base
Based on the transverse distribution curve of Earth pressure along the base, the stress variation characteristics of the foundation at different transverse positions during the loading filling process of the upper embankment could be investigated; thereby the design parameters of composite foundation could be optimized [30–33]. Figure 3 shows the curve of the Earth pressure distribution along the base in the natural foundation and in the pile-net composite foundation. Figure 3(a) suggests that the Earth pressure gradually increases with the increasing filling load. The testing values and measured values are becoming closer in natural foundation, but there is a large difference in the composite foundation because it is affected by the difference of pile-soil objective stiffness. The Earth pressure on top of the piles is higher than that between the piles and the Earth pressure along the base shows a “jagged” distribution. Meanwhile, there are differences in the measured values of symmetrical measuring points, resulting from the effects of temporary carriageway load at both sides of the road shoulder and pile foundation quality. The effect of weather can also lead to the differences in the measured values of symmetrical measuring points because of the difference in temperature, humidity, and the accumulation of water.

(a)

(b)
The variation of soil pressure in the pile-net composite foundation is closely related to geological conditions and design parameters of composite foundation. Xu et al. [20] have analyzed the load transfer mechanism of pile-net composite in soft base with high strength, where soft ultimate longitudinal and transverse tensile strength is larger than 300 kN/m. It could be found that the soil pressure of pile-net composite foundation with medium and low compressibility is similar to that of soft soil with high compression when time and filling load are increasing. Meanwhile, compared with the medium and low compressibility soil, the soil pressure of the composite foundation in the soft soil area with high compressibility is more sensitive to the filling load.
3.3. Soil Arch Effect Analysis in Foundation Filling
Because of the difference in stiffness between piles and soil between piles, there exists a difference in soil settlement between the soil on the top of pile and that between piles under the soil weight, and the settlement of soil between piles is greater than that of soil on the top pf pile. Due to the redistribution of stress caused by the change of soil displacement, the direction of large principal stress is deflected, resulting in the compaction of soil in the circular arch region of adjacent piles. Therefore, a compacted shell arch with slightly higher stiffness is formed, and the upper load is redistributed to the pile top [34, 35]. In the theoretical analysis of the soil arch effect, the stress reduction ratio is usually used to quantify the decrease in the vertical stress caused by the soil arch effect as follows:where σs is the stress between the piles, γ is the weight of soil, h is the height of embankment, and Sr is the stress reduction ratio. When Sr < 1, the soil arch begins to work and when the value is close to 0, it works better.
Figure 4 shows the relationship between Sr and γh in the composite foundation. It could be found that the stress reduction ratio gradually decreases with increasing γh, and the soil arch is gradually formed. When γh is more than 80 kPa, Sr is less than 1, and the soil arch effect begins to work. When the foundation filling is completed, Sr is approximately 0.6.

3.4. Change in the Coefficient of Subgrade Reaction in Foundation Filling
According to the soil pressure in the field test, there exist upper filling loads between the pile and the soil. In some cases, the Winkler model can be used to simulate the opposing force between the pile and the soil [36, 37]. One unit in the foundation is selected and the coefficient of subgrade reaction is as follows:where K is the coefficient of subgrade reaction, σs is the soil pressure, and δs is the settlement between the pile and soil.
As shown in Figure 5, there is a positive correlation between σs and δs. As the settlement increases, the average soil pressure increases at a fast rate at first before increasing slowly. Meanwhile, during the foundation filling period, according to the curve of K and δs/(s-d) in Figure 6, K shows a downward trend when the settlement increases. It could be found that K experiences a large decrease during the loading period but a relatively small decrease during the settlement period, and then it almost becomes stabilized. Finally, K remains unchanged at 6.4 kN/m3 in the stabilized period.


3.5. Analysis of Pile-Soil Stress Ratio
Pile-soil stress ratio is one of the important reference indexes for pile-net composite foundation design, which refers to the ratio of the average stress on the pile top to the average stress on the soil surface between the piles [38, 39]. Figure 7 suggests that the pile-soil stress ratio increases gradually over time. In the beginning period, the stress ratio is small, but it has a relatively large increase during the whole filling period and becomes stabilized during the settlement period. The pile-soil stress ratios in the road shoulder and the center are 3.2 and 2.8, respectively. The change mechanism of pile-soil stress ratio is analyzed. The elastic modulus of pile is much larger than that of soft soil foundation, leading to the pile bearing more upper load in the loading process. As the differential settlement of pile and soil increases, the trend of load bearing differentiation is further strengthened. When the filling load does not continue to increase, the stress ratio tends to be stable.

Compared with the results from other researches [40–42], the pile-soil stress radio at stable stage in this project is still relatively small, so improved methods could be applied to further increase pile-soil stress radio. Researches have shown that the stiffened cushion can transfer the upper load to the pile top, resulting in the rise of pile-soil stress ratio, thus improving the bearing capacity of the foundation. Furthermore, the effect of three-direction geogrid on the pile-soil stress ratio is better than that of bidirectional geogrid and unidirectional geogrid, and the setting of pile cap on the pile top has a very significant effect on the pile-soil stress ratio when the reinforcing material of cushion is three-direction geogrid.
3.6. Development of Excess Pore Water Pressure
The influence of pile on the surrounding foundation soil under vertical load could be regarded as the stress of foundation soil when vertical concentrated force is applied to the interior of semi-infinite body. When the filling load increases rapidly, the soil of the same volume as the pile extrudes outward, and the soil around the pile is disturbed by strong compaction; thereby the soil structure is destroyed. Meanwhile, a remodeling zone of a certain thickness is formed around the pile, and the instantaneous excess pore water pressure is generated. The formation and development of excess pore water stress reduce the effective stress of soil, resulting in the change of soil deformation and strength, so it is necessary to analyze the development of excess pore water pressure in the short pile-net composite foundation project [43–45].
In the road filling period, the excess pore water pressure increases with the filling load and with some hysteresis qualities, as shown in Figure 8. In this period, the change in the pore water pressure is mainly influenced by the additional stress from the loading [45, 46]. In the settlement period, the foundation pore water pressure gradually disappears and becomes stable because the filling load does not change. At WP-4-1, the pore water pressure changes slightly, but it has a large change in the lower layers. Therefore, it proves that the short piles could better pass additional stress from the upper fill layers to the soil layers.

4. Summary and Conclusions
In this study, Gan-Long high-speed railway project is chosen as the research object. In situ experiments and laboratory experiments are combined to obtain the physical and mechanical characteristics of the soil in the test section. Then, the mechanism of load transmission in the upper foundation in the short pile-net composite foundation of medium compacted soil is studied. The main conclusions are as follows:(1)The results of in situ and laboratory experiments show that the physical characteristics of the different layers in the test section are similar. The natural moisture content is approximately 24%, the void ratio is approximately 0.7, and the liquid limit is approximately 40%, which has the characteristics of low moisture content, low void ratio, and high liquid limit. The compressibility ratio is relatively stable at approximately 0.25, which belongs to medium compressible soil.(2)The theoretical and test values of the transverse distribution of the natural foundation Earth pressure along the foundation are very close. In the composite foundation, the test values on the top of the piles are almost 2.5 times higher than those between the piles. The Earth pressure in the central foundation has an error of approximately 6.67% between the test value and the theoretical value. In a composite foundation, the Earth pressure distribution along the base shows “jagged” characteristics.(3)The soil arch begins to work when the filling load reaches 80 kPa. The coefficient of the subgrade reaction decreases with the increasing foundation settlement, and the value is 6.4 kN/m3 when the foundation is stable. The stress ratio in the road shoulder and the center is 3.2 and 2.8, respectively. The change of pore pressure in the pile top test section is higher than that in the reinforcing area, which proves that short piles can better pass the additional stress from the upper layers to the soil layers.
In this study, the pile cap and high-strength grille have not been applied in the field test, so the bearing characteristics of the short pile-net composite foundation using pile cap and high-strength grille can be further studied in the future. Meanwhile, the characteristics of soil lateral displacement can also be analyzed to verify the restraint of pile on soil lateral displacement.
Data Availability
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments
This study was funded partially by the Science and Technology Project of Zhejiang Provincial Department of Transportation, Grant no. 2021044.