Evaluation of code provisions for seismic performance of unachored liquid storage tanks

Abstract

Seismic performance of two unanchored liquid-storage tanks with tank diameter of 24.5 m and 36 m and operating liquid height of 12.2 m and 20.0 m, respectively were investigated using Coupled Eulerian-Lagrangian (CEL) and mechanical spring-mass analogy nonlinear finite element computational methods. The CEL approach includes the effects of higher modes of liquid vibration (sloshing), liquid breaking effects, and liquid-structure interaction during seismic loading. The modern seismic design provisions for liquid-storage tanks, on the other hand, are based on a mechanical spring-mass analogy. This approach neglects the higher vibration modes for the sloshing water, liquid-structure interaction, and effects of tank base uplift on seismic performance. For the tanks, base uplift histories were computed with both modeling approaches through nonlinear time history analysis performed using five recorded earthquake acceleration data. The uplift histories were compared to evaluate the adequacy of code seismic design provisions for unanchored tanks, and to determine whether the mechanical spring-mass analogy can be used to predict seismic performance of unanchored tanks. Analysis results show that the traditional mechanical spring-mass analogy, which is the basis for the current seismic design provisions, does not capture tank uplift history and its effects on dynamic loads. This approach underpredicts the total numbers of tank uplifts during seismic loading. The maximum tank base uplift computed using mechanical spring-mass analogy had an average error between 22% and 58 % for each tank. The results show that there is a need to developed a modify version of the traditional mechanical spring-mass analogy to be used for predicting seismic performance of unanchored liquid-storage tanks.