Design And Construction of A 9-Meter-High Embankment Over Champlain Sea Clay

ABSTRACT
Construction bearing on normally, or minimally over-consolidated, sensitive clays of the Champlain Sea deposit in Eastern Canada has traditionally forced the implementation of diverse engineered solutions. The present paper reflects on a case study where a 9-meter-high embankment required a comprehensive geotechnical investigation campaign, and the ensuing engineered solutions. The embankment, located in the Canadian National Capital Region, is part of a multi-billion-dollar transit project executed using a fast-paced P3 delivery model. As such, firstly the project required a particular emphasis on schedule optimization. Secondly, a right of way constraint on the northern face entailed the implementation of a retaining structure to limit the footprint of the embankment. Finally, a contractual performance criterion dictated that total post construction settlements, and deflections during operations be limited to under 5mm.

The subsurface encountered at the site mainly consisting of a silty clay deposit underlain by Till and resting on bedrock, was subjected to a multi-phased investigation campaign. Though the baseline geotechnical data provided by the client suggested that the silty clay deposit’s OCR ranged between 3.4 to 4.5, supplemental investigation carried out by AtkinsRéalis identified a much lower OCR of about 1.9 – which translated to a governing over-consolidation margin of about 60 kPa. Following a series of analyses and design iterations the retained design solution was able to address the project constrains. This included the implementation of accelerated consolidation of the silty clay deposit using vertical wick-drains and surcharging to meet tight deadlines. Consequently, the resulting ground improvement through consolidation strength gains provided adequate resistance to support the retaining structure. And finally, the targeted use of lightweight EPS embankment fill allowed the mitigation of settlements to adhere to the strict threshold. This was supported by data from a network of geotechnical instrumentation and monitoring systems with near perfect correlations with predicted behaviors.