UC Berkeley

There have been numerous examples of damage to buildings due to the effects of soil liquefaction or cyclic softening in recent significant earthquake events. Of these examples, shallow-founded buildings on level ground atop shallow liquefiable soils have often been impacted by partial or complete bearing failures, loss of foundation soils due to eroded sediment ejecta, and vertical settlements that have been exacerbated by soil-structure-interaction (SSI). While significant progress has been made in recent years towards understanding these deformation mechanisms, there is still no widely-accepted simplified method by which engineering practitioners can reliably estimate the settlement of buildings due to liquefaction or cyclic softening. Rollins and Seed (1990) pointed out that the liquefaction potential below a building was often evaluated by treating the soil as if it were in the free-field. Two decades later, Bray and Dashti (2010) found that liquefaction-induced building settlements are still often estimated using empirical procedures developed to calculate post-liquefaction, one-dimensional, consolidation settlements in the free-field (e.g., Tokimatsu and Seed, 1987; Ishihara and Yoshimine, 1992). These free-field procedures do not capture important shear-induced and localized volumetric-induced building deformation mechanisms. Therefore, these procedures can significantly underestimate building settlements. Improvements are required to advance the state-of-the-art in liquefaction engineering.

The primary objective of this research effort is to advance the understanding of the seismic performance of buildings subjected to soil liquefaction or cyclic softening. This objective is achieved through the performance of a geotechnical centrifuge experiment and through the documentation and interpretation of a number of field performance case histories. Well-documented field and physical model case history data are essential to advancing the profession's understanding of the effects of liquefaction or cyclic softening on building performance. The field and laboratory studies are also critical to the development and calibration of simplified empirical procedures and advanced analytical procedures. Therefore, while the primary objective of the work presented herein is to advance the understanding of liquefaction-induced building settlements, an additional significant objective is to provide high quality and well-interpreted field and physical model case history data that can be used to develop, evaluate, and calibrate analytical procedures and design methods.

A sophisticated geotechnical centrifuge model of shallow-founded model structures atop shallow liquefiable soils was designed, constructed, and subjected to several realistic earthquake motions of varying intensity and duration as part of this work. This experiment was designed to build on previous centrifuge experiments described by Hausler (2002) and Dashti (2009). There were many design similarities between this model and the models described by Dashti (2009), and the findings were generally consistent. However, an additional key objective of this model was to evaluate the effects of building adjacency on the observed performance of the model structures. It was found that placing a baseline model structure next to an identical model structure did not significantly impact the building settlement of either building during the test. Placing the baseline model structure next to a taller, heavier shallow-founded model structure illustrated that it is erroneous to always expect a heavier shallow-founded structure to settle more than a lighter shallow-founded structure when subjected to shallow liquefaction. For some ground motions, less cyclic softening occurred under the heavier structure, so it settled less than the adjacent lighter baseline structure. The seismic demand on the building's foundation soils is directly related to the ground motion and both the building's weight and its dynamic response. In addition, the adjacent model buildings generally tilted and displaced laterally away from one another. These observations suggest that the physical presence of the adjacent building kinematically constrained ground movements under the structures on the sides nearest the adjacent buildings relative to the ground movements that occurred under the sides away from the adjacent building. Also, the confining stresses under the sides of closely-spaced adjacent buildings are lower under the side away from the neighboring building than the side closest to the neighboring building. Consequently, with all other things equal, one would expect a lower liquefaction resistance in the foundation soils on the sides away from the adjacent building, and therefore, relatively more cyclically-induced deformation to occur on that side during shaking

Building performance evaluations of select buildings in Christchurch, New Zealand during the 2010-2011 Canterbury earthquake sequence were also performed. Many multi-story buildings were heavily damaged by liquefaction-induced ground movements during the 22 February 2011 Mw 6.2 Christchurch earthquake, but not by other significant earthquakes. Conventional CPT-based liquefaction triggering evaluations were generally conservative for these field case histories, and this conservatism led to post-liquefaction free-field ground settlement estimates that were generally similar for the 4 September 2010 Mw 7.1 Darfield, 13 June 2011 Mw 6.0, and 22 February 2011 Mw 6.2 Christchurch earthquakes. Variability in the shallow subsurface conditions over relatively short distances was sometimes a critical factor in the seismic performance of the ground and buildings. Ground loss due to eroded sediment ejecta was found to be an important foundation settlement mechanism in several cases of shallow-founded buildings with shallow liquefiable foundation soils. In addition, two shallow-founded, tall buildings settled and tilted due to relatively deeper liquefaction or cyclic softening during the Christchurch event. Though these tall structures were founded at least 3 m above materials that were judged to have liquefied, the liquefied materials were well within the depth range with a significant vertical strain influence using the settlement approach developed by Schmertmann (1978) for rigid footings on sand. Therefore, cyclically-induced deformation of these materials due to increased pore water pressures likely led to the observed building settlements and tilts.

In summary, there are important physical constraints caused by building adjacency that can affect the seismic performance of buildings subjected to soil liquefaction or cyclic softening. The performance of a building is also significantly affected by the earthquake motion, the ground beneath its foundation, and its dynamic response, including SSI effects such as superstructure rocking. Consequently, the combined effects of all these factors are difficult to judge a priori. Additional studies are warranted to develop insights and improved methods.