Liquefaction, Lateral Stress, Consolidation State, and Aging

Abstract

In 1986–1989, Jackson Lake Dam, Wyoming, was removed after standard penetration tests indicated that the foundation soil was subject to liquefaction. Part of a subsequent research program involved measuring lateral in situ stresses at various depths in the foundation soil by use of the Ko Stepped Blade, a device that introduces a stepwise lateral displacement of soil so contact pressures can be extrapolated to give the hypothetical pressure on a zero-thickness step. The foundation soil was cohesionless mix of alluvial and deltaic silts and sands with some sharp-edged gravel. The lateral stress measurements established relationships between the four topics in the title of this paper. A normal consolidation state, meaning that a soil is in equilibrium with its own weight, can be established when Ko is constant with depth and lateral effective stresses extrapolate to zero at the ground surface. On this basis, three stress zones were identified: (1) Based on limited data, an uppermost zone, to a depth of 4 m (13 ft) underneath the dam, was identified as being normally consolidated under the weight of the dam. This is as expected and probably occurred during construction. Underneath the upper zone was (2) a transition zone, 4 to 5.5 m (13 to18 ft) in depth, and under that (3) a lower zone, 5.5 to 10 m (18 to 33 ft) in depth. The lower zone was determined to be normally consolidated without the weight of the dam. Ko without the weight from the dam was determined to be 0.5, and lateral effective stresses extrapolate to zero at the predam ground surface. Note that the value of Ko was measured and not derived from empirical correlations. Nevertheless, it is consistent with normal consolidation and a friction angle of 30 degrees reported for this deposit. Despite having been characterized as cohesionless, the soil deep underneath the dam had resisted consolidation for over 60 years, and the depth to the underconsolidated layer is consistent with observations that liquefaction can be deep-seated. The existence of such a layer calls into question the use of a temporary surcharge for site preparation for cohesionless soils, or a reliance on measurements that settlement has stopped. A similar situation can be anticipated under existing structures such as dams and highway embankments that are located on the geologically young but historically old granular soil deposits of floodplains and deltas. The weight of the dam reduced Ko in the lower zone soil from 0.5 to 0.35. This can happen only if the soil has developed sufficient strength to resist consolidation and can be attributed to aging. Mohr circles drawn from the lateral stress measurements indicate a gain in cohesion of the order of 34 kPa (4.9 psi or 700 psf) from aging. The gain also may be explained by an increase in the friction angle from dilatancy attributed to cementation rims at grain contacts. The site was treated by deep dynamic compaction that increased Ko to 1.0, but with considerable variability. That value indicates a temporary liquid condition, and may have been preserved by seepage forces when pore water pressure supported the entire weight of the soil column. The replacement dam reduced the average Ko from 1.0 to about 0.5, which should be sufficient prevent future liquefaction. However, the cap on Ko of 1.0 during ramming can restrict the amount of additional load if the final Ko is to be no lower than the normal consolidation value.