Large earthquakes in the last 25 years have caused significant damage to buildings and infrastructure, including the partial or total collapse of storage tanks in various industries. Elephant foot buckling, or local buckling at the base, is one of the main failure modes observed in these structures, and this failure mode can lead to their collapse and/or complete loss of contents. Although hydrostatic and hydrodynamic loads typically affect the seismic response of tanks, the effect of soil type on tank buckling behavior has not been widely studied or recognized. This research aims to evaluate the effect of soil type on seismic fragility of tanks by analyzing typical storage tanks used in the wine industry. The work focuses on elephant foot buckling for tanks with both unanchored and anchored bases and compares the influence of three different types of soil and two different tank geometries. The approach uses the capacity spectrum method, as opposed to the more commonly used incremental dynamic analysis, to determine a critical peak ground acceleration to cause buckling at the tank. The tanks were subjected to 21 Chilean seismic records with three different soil types and a no-soil condition. From the results a lognormal fragility curve, and its median and standard deviation, are calculated. The results indicate that unanchored tanks built softer soils exhibit poorer performance, while tanks in competent soils and rock exhibit good performance. Anchored tanks show less sensitivity to soil types than unanchored tanks. The study demonstrates the importance of considering soil-foundation-structure interaction for wine storage tanks, but the results indicate that many comparable storage structures will be similarly affected.

In the last decades, numerous liquid storage tanks have been affected by strong earthquakes, the damage observed ranges from the partial collapse to the total collapse of the storage tanks. Elephant-foot buckling is one of the most common failures observed in these structures, which can provoke their collapse and complete loss of contents. While hydrostatic and hydrodynamic loads typically impact the seismic response of tanks, the soil type on which they are built plays an important role in influencing their performance during earthquakes. However, the soil-tank interaction has not been considered in the seismic fragility analyses of continuously supported tanks. This research aims to evaluate the seismic fragility of a continuously supported wine storage tank with a particular focus on elephant-foot buckling considering the soil-tank interaction. A specific soil condition and a typical wine storage tank are evaluated utilizing pushover-based seismic analysis and the Capacity Spectrum Method (CSM). 3D nonlinear Finite Element (FE) models are developed considering the tank, foundation, and soil. Seven ground motion records compatible with the soil type are considered. The seismic fragility is estimated using the FE models and the ground motion records. Both unanchored and anchored conditions are evaluated. The obtained results show that for the considered case study, the anchored condition shows better seismic performance when compared to the unanchored condition.

This research addresses experimentally the relationship between the excitation frequency and both hoop and axial wall stresses in a water storage tank. A low-density polyethylene tank with six different aspect ratios (water level to tank radius) was tested using a shake table. A laminar box with sand represents a soil site to simulate Soil-Structure Interaction (SSI). Sine excitations with eight frequencies that cover the first free vibration frequency of the tank-water system were applied. Additionally, Ricker wavelet excitations of two different dominant frequencies were considered. The maximum stresses are compared with those using a nonlinear elastic spring-mass model. The results reveal that the coincidence between the excitation frequency and the free-vibration frequency of the soil-tank-water system increases the sloshing intensity and the rigid -like body motion of the system, amplifying the stress development considerably. The relationship between the excitation frequency and wall stresses is nonlinear and depends simultaneously on both sloshing and uplift. In most cases, the maximum stresses using the nonlinear elastic spring-mass model agree with those from the experiments.