Soil instability from tsunami hazards causes substantial damage to coastal infrastructure (e.g., the damage caused by the 2011 Great East Japan Tsunami, or the Heisei Tsunami). Tsunamis are unpredictable, so it is difficult to obtain field-scale measurements. Simulating tsunamis in a laboratory setting is therefore important to further understanding of soil instability induced by tsunami inundation processes. Such laboratory simulations are difficult because the scale of tsunami inundation processes is very large, hence dynamic similitude cannot be achieved for small-scale models. The ability to control the body force in a centrifuge environment considerably reduces the mismatch in dynamic similitude.
A novel centrifuge apparatus specifically designed for exploring the physics of soil response to tsunami-like loading was designed, constructed, and customized for use in the 9.1 m geotechnical centrifuge at the University of California, Davis. Tsunami flooding in the apparatus is created by lifting a gate and releasing water from a reservoir (i.e. breaking of a dam), and tsunami drawdown is achieved by lifting another gate to drain the flooded water. The ability to initiate both runup and drawdown in a single experimental run makes this device the first of its kind. Flow measurements indicate that the apparatus can produce tsunami-like loading with sufficiently high water pressure and flow velocities. More specifically, the present 1:40 model represents the equivalent geometric scale of the prototype soil field of 9.6 m deep, 21 m long, and 14.6 m wide, and the pressures and flow velocities in the model are identical to those of the prototype, which yields realistic condition of flow-soil interactions. There is no laboratory facility capable of creating such laboratory conditions under the normal gravitational condition.
Using the apparatus, the soil response from two experiment configurations is analyzed and presented. The configurations are: (1) a homogeneous soil specimen, and (2) a homogeneous soil with the inclusion of an embedded impermeable layer. Scour is observed during tsunami runup and 3D laser scans of the soil surface before and after the experiment show the volume of soil removed. Digital image processing is used to characterize the tsunami runup and drawdown flow conditions. The soil’s pore water pressure response to the fast-moving surging front is observed, as well as to the release of water during the tsunami drawdown stage. In contrast to previous computational models of soil response to the tsunami runup process, our results show that pore water pressure increases ahead of the tsunami surge front, showing a significant effect of lateral pore pressure gradients. The increase in pore water pressure ahead of the surge front results in upward pressure gradients reducing the soil stability. The pore water pressure results of the two experiment configurations are also compared. It is found that the inclusion of an embedded impermeable layer significantly changes the pore pressure patterns in the model from those observed in the homogeneous soil specimen. Additionally, the vertical effective stress beneath the impermeable layer is reduced more than 50% during drawdown.