Shake table lateral earth pressure testing with dense c-ϕ backfill

• A dynamic large shake table experiment was conducted for a dense backfill with cohesion and friction.
• Before limit equilibrium earth pressure occurs, backfill vibration and wall movement control the lateral force.
• Limit equilibrium earth pressure force matches analytical estimates that include the soil cohesion.
• All recorded experimental data sets are readily available at the NEES data repository.

Dynamic lateral earth pressure is recorded during ten shake table testing events. In these tests, peak input acceleration at the base of the retaining wall varied in the wide range of 0.13–1.20 g in order to include scenarios of relevance to recently observed strong earthquake excitations. The results shed light on the influence of soil cohesion and the effect of small wall movements on the magnitude and distribution of earth pressure. In accordance with field practice, a commonly encountered dense sand backfill with a small percentage of fines (SP-SM) is used. Inside a large soil container, earth pressure is measured against a rigid wall (backfill height H=1.7 m) that is allowed to undergo limited translation/rotation due to the imparted dynamic excitation (up to 10 mm or 0.006H at 1.2 g base acceleration). In this particular series of experiments, favorably low dynamic pressures were recorded at backfill accelerations of up to about 0.7 g in light of: (i) the relatively high soil strength (including cohesion) that precluded a limit equilibrium type failure in the backfill, and (ii) the high soil stiffness coupled with the small value of observed wall translation/rotation (as much as 3 mm or 0.0018H at ground surface). In tests with instants of very high acceleration (in the range of 1 g), the corresponding dynamic earth pressure is found to be of much significance for practical applications. Lateral thrusts recorded during these instants of strong shaking compare well with limit equilibrium predictions that include the soil cohesion intercept. Exclusion of the cohesion intercept results in substantial over-prediction of the measured lateral forces.