Column-supported embankments with or without basal geosynthetic reinforcement can be used in soft ground conditions to reduce settlement by transferring the embankment load to the columns through stress redistribution above and below the foundation subgrade level. Column-supported embankments are typically used to accelerate construction and/or protect adjacent facilities from additional settlement. The column elements consist of driven piles or formed-in-place columns that are installed in an array to support a bridging layer or load transfer platform. The bridging layer is constructed to enhance load transfer using several feet of compacted sand or sand and gravel that may include one or more layers of high-strength geotextile or geogrid reinforcement.

Mobilization of the mechanisms of load transfer in a column-supported embankment requires some amount of differential settlement between the columns and the embankment as well as between the columns and the foundation soil. When the embankment height is low relative to the clear spacing between columns, there is the risk of poor ride quality due to the reflection of the differential foundation settlement at the surface of the embankment. The minimum embankment height where differential surface settlement does not occur for a particular width and spacing of column is the critical height. The conventional approach is to express critical height as a fixed ratio of the clear span between adjacent columns; however, there is no consensus on what ratio to use and whether a single ratio is applicable to all realistic column arrangements.

The primary objective of this research is to improve the understanding of how column-supported embankments deform in response to differential foundation settlement. A bench-scale experimental apparatus was constructed and the equipment, materials, instrumentation, and test procedures are described. The apparatus was able to precisely measure the deformation occurring at the sample surface in response to differential settlement at the base of the sample. Critical heights were determined for five combinations of column diameter and spacing representing a wide range of possible column arrangements. In addition, tests were performed using four different column diameters in a single column configuration with ability to measure the load acting on the column and apply a surcharge pressure to the sample. In total, 183 bench-scale tests were performed over a range of sample heights, sample densities, and reinforcement stiffnesses. Three-dimensional numerical analyses were conducted to model the experiments. The critical heights calculated using the numerical model agreed with the experimental results.

The results of the laboratory tests and numerical analyses indicate that critical height depends on the width and spacing of the columns and is not significantly influenced by the density of the embankment fill or the presence of reinforcement. A new method to estimate critical height was developed and validated against extensive case histories as well as experimental studies and numerical analyses performed by others.