Hybrid Concrete-and-Cemented-Soil Piles Tame Deep Soft Soils Beneath Highway Embankments

Beneath the endless ribbons of highway that cross China’s eastern coastal plains lies a problem as old as civil engineering itself: soft, waterlogged soil that squirms under load. When engineers on the Jiangsu expansion of the Beijing–Shanghai Expressway needed to build a three-meter-high embankment across deep, compressible mud, they turned to an increasingly celebrated hybrid technology known as the PCCS composite foundation—a precast concrete pile inserted into a surrounding column of cemented soil. A new field-and-simulation study published in Case Studies in Construction Materials now offers the most detailed picture yet of how these long-core composite piles share the burden of an embankment, and the numbers are striking: the piles ultimately carried roughly two-thirds of the entire embankment load while holding settlement to a few centimeters.

The PCCS system, short for precast concrete piles reinforced by cemented soil, marries two established ground-improvement techniques. A deep-mixing column, created by blending cement into soft soil in place, forms a stiff outer sleeve. Into that sleeve, engineers insert a slender precast concrete pile that acts as a rigid inner core. The two components work in complementary fashion: the concrete core supplies axial stiffness and bearing capacity, while the cemented-soil column enlarges the effective diameter of the pile, expands the contact area with the surrounding ground, and improves lateral confinement and interfacial bonding. Depending on whether the inner core is shorter, equal to, or longer than the outer column, the system is classified as short-core, equal-core, or long-core. For deep and thick soft-soil deposits—the kind encountered along the Beijing–Shanghai corridor—only the long-core variant makes sense, because the extended concrete core can reach down toward a competent bearing stratum that a shallow cemented column alone could never touch.

The field test section sits between chainages K950+054 and K950+158 of the expressway expansion. The original design called for conventional precast concrete piles of 28 to 35 meters, but the project team adopted the PCCS alternative for its combination of high bearing capacity and cost savings. Each PCCS unit consisted of a 32-meter-long, 0.4-meter-diameter precast concrete core of C60-grade concrete topped with a 1.6-meter square pile cap, surrounded by a 12-meter-long, 0.8-meter-diameter cemented-soil outer core. The piles were arranged in a square grid at 3.5-meter center-to-center spacing, and a 0.3-meter gravel cushion laced with a geocell—a honeycomb-like polymer reinforcement—was laid over the pile heads to distribute load and limit differential movement at the ground surface.

To dissect the load-transfer mechanics, the research team built two fully coupled fluid–solid numerical models in ABAQUS: a full half-width cross-section of the embankment and an idealized unit cell centered on a single pile. The deep saturated soft soil was represented with the modified Cam-Clay constitutive model, its compression and swelling indices calibrated against laboratory consolidation and triaxial tests, while the embankment fill, cushion, and cemented soil used the Mohr-Coulomb elastoplastic model. A bond–slip friction model captured the interface between the concrete core and its cemented sleeve, with a bond strength of 200 kilopascals and a tangential friction coefficient of 0.65 drawn from prior experimental work. Mesh sensitivity checks confirmed that predictions of settlement and pile–soil stress ratio changed by less than 2 percent between medium and fine meshes, giving the team confidence in the computational framework.

Validation against field instrumentation proved convincing. At 320 days after construction, measured pile-top stress fell between the two models, with the unit model deviating by at most 12 percent from observations, and normalized settlement profiles along the embankment height agreed within 4.2 percent. The simulations revealed a clear temporal choreography: during embankment filling, both pile and soil stresses climbed steadily, but once construction ended their paths diverged. The pile-top stress kept rising as the soft ground consolidated and drained, while the stress in the surrounding soil peaked and then gradually declined—signature behavior of soil arching, the process by which load sheds from yielding soil onto stiffer inclusions through intergranular shear. In the final stabilized state, the surrounding soil carried only about 33 kilopascals, roughly 55 percent of the embankment’s average applied pressure, while the pile top reached 266.2 kilopascals, yielding a pile–soil stress ratio of 8.1. In load-sharing terms, the PCCS carried 68 percent of the total embankment weight and the soil just 32 percent.

Settlement performance was equally impressive. Immediately after filling, the pile and the adjacent soil had settled 6.9 and 22.5 millimeters respectively; as consolidation proceeded, the soil continued to compress, stabilizing after 204 days at a final measured settlement of 47.5 millimeters. The pile top, by contrast, essentially stopped moving about 86 days after construction, finishing at just 15.2 millimeters. The growing gap between pile and soil settlement—eventually stabilizing at 31.3 millimeters after 264 days—is precisely what drives soil arching above the pile caps. Numerical stress contours captured the result vividly: as the embankment rose, principal stress concentrated above the pile tops in a distinctive mushroom-shaped pattern, and stress vectors tilted outward to form a coherent arch. Once the arch was fully developed, the upper embankment fill became nearly differential-settlement-free, meaning the pavement above rode on a smoothly supported platform despite the uneven foundation below.

The study quantified this arching with a critical height of approximately 1.15 times the net pile spacing—the distance between pile-cap edges. Below that plane, differential settlement and stress redistribution are concentrated; above it, settlement curves converge and the embankment behaves as a uniform body. The analysis also distinguished between an inner arch, spanning adjacent piles, and a higher outer arch spanning the diagonals of the pile grid. At an embankment height of 5 meters, the outer arch reached a maximum of 1.55 times the net spacing above the embankment base, reflecting the wider diagonal pile spacing. The horizontal earth pressure coefficient, a direct measure of arching intensity, stayed near the at-rest value through most of the embankment height but surged near the base, where at 5 meters of fill it reached values 3.2 and 2.3 times those measured under 3-meter and 4-meter embankments respectively.

Parametric sweeps explored how the embankment fill’s own properties modulate the arch. Increasing the fill’s elastic modulus modestly reduced peak stress and displacement—each 10-megapascal increment trimmed normalized stress by roughly 4 to 5 percent and displacement by about 3 percent. Cohesion proved more influential: every 5-kilopascal gain cut maximum normalized displacement by 7 to 11 percent, indicating that a more cohesive fill helps the arch develop and hold its shape. The internal friction angle, surprisingly, mattered little, with stress and displacement fluctuating within 3 percent for each 5-degree change. The geocell embedded in the cushion added its own membrane effect, deforming symmetrically along the embankment axis and developing peak tensile force near the shoulder pile, where it helped distribute lateral load and maintain stability at the slope.

Horizontal ground movements told a consistent story. Displacements decreased with depth but showed a localized bulge between 4 and 12 meters during the first 120 days after construction, and intensified from the surface down to 16 meters as time passed. At the slope toe, the computed maximum horizontal displacement was 4.4 millimeters at the end of construction against 2.5 millimeters measured in the field; by 365 days the computed value reached 14.9 millimeters, exceeding the field measurement by only 15.5 percent. Axial-force profiles along the pile depth revealed a neutral point at 16 meters for the central pile, below which negative skin friction gave way to positive resistance, while piles beneath the slope, carrying less overburden, showed neutral points at 9.0 and 4.0 meters for their inner and outer cores respectively.

The authors are candid about the model’s limits: pile installation disturbance, long-term creep of the soft clay, cemented-soil degradation, and progressive interface damage were not explicitly simulated, and these simplifications likely explain why measured soil settlement slightly exceeded predictions. Still, the agreement achieved across stress, settlement, and displacement metrics demonstrates that the long-core PCCS is a genuinely effective answer to deep soft-soil embankments—delivering rigid-pile settlement control at a fraction of the cost of conventional piling, and giving engineers a quantified, field-validated framework for designing the next generation of highways across the world’s soft deltaic plains.

Subject of Research: Load transfer mechanism and settlement performance of long-core PCCS composite foundations supporting embankments over deep soft soil

Article Title: Load transfer mechanism and performance evaluation of the PCCS composite foundation with long-core under embankment load

Article References: Zhang, C., Zhang, D., Liu, S., Wang, Z., Liu, L., & Zhu, M. (2026). Load transfer mechanism and performance evaluation of the PCCS composite foundation with long-core under embankment load. Case Studies in Construction Materials, 25, Article e06506. https://doi.org/10.1016/j.cscm.2026.e06506

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06506

Keywords: PCCS composite foundation, precast concrete pile, cemented soil, soil arching, embankment, soft soil, pile-soil stress ratio, settlement control, geocell reinforcement, modified Cam-Clay model, Beijing-Shanghai Expressway, ground improvement