No abstract

Many highway embankments experience problematic settlements. Compression of soil under the self-weight of the embankment generally occurs during construction, but postconstruction, wetting-induced collapse can result in more long-term settlement problems, depending on climate and movement of the wetting front. A study undertaken to examine settlement associated with unsaturated soil embankments included centrifuge modeling of compacted silt embankments, laboratory testing of the embankment soil, and data interpretation with the focus on settlement prediction. A silty soil was selected to facilitate the measurement of matric suction using tensiometers. Three model embankments were constructed and tested in the Army Corps of Engineers Centrifuge Research Center in Vicksburg, Mississippi. Embankments 20 m high were simulated using centrifugal acceleration of 165 g. Embankments were constructed to achieve a relative compaction of 90 or 95 percent based on standard effort and moisture content between 2 and 5 percent dry of the optimum moisture content. Instrumentation used during self-weight compression and wetting included linear variable differential transformers and pore-pressure transducers equipped with high-air-entry porous stones. Results demonstrate the importance of the as-compacted water content and dry unit weight on the potential for wetting-induced collapse settlement. Settlement caused by self-weight compression and that caused by wetting-induced collapse are clearly discernable in results of centrifuge tests, allowing for comparison to settlement predictions. Results are discussed in light of typical compaction specifications, oedometer-based predictions, and implications for the design of compacted embankments.