Numerous unanchored liquid storage tanks exist in the field serving vital roles in our daily life. They are important components of lifeline and industrial facilities, critical elements in municipal water supply and fire fighting systems, and used in many industrial facilities for storage of water, oil, chemicals and liquefied natural gas. Most of the tanks in the field are unanchored because they are easier and cheaper to construct. Anchored tanks must be connected to large expensive foundations to prevent tank uplift in the event of earthquake occurrence. In contrary, unanchored tanks are constructed above compacted soil with cheap ring wall foundation underneath of the tank wall, though sometimes ring walls are omitted.
Current design standards are inconsistent in predicting the behavior of unanchored tanks. In general, the estimation of liquid-exerted overturning moments lacks explicit consideration of the support effects, and may not even include shell flexibility and site conditions. The capacity of the bottom plate to resist uplifting forces is generally underestimated because membrane forces are neglected. In addition, the allowable buckling stress in the shell of unanchored tanks lacks the consideration of stress concentration resulting from base plate uplift, foundation flexibility and tank radial deformation. The large scale failure to unanchored tanks during recent earthquakes highlighted the need for a more accurate design standard for such tanks. The aim of this study is to employ the sophisticated finite element analysis technique, developed by the proposal investigators and funded by grants from NSF and NCEER , to come up with simplified but accurate design guidelines for unanchored liquid storage tanks.
A literature review of available design standards shows that the API 650  standard for oil tanks (or alternatively, the AWWA  standard for water tanks) have been the most commonly used standards for seismic design of tanks. In recent years, the New Zealand recommendations for seismic design of storage tanks  have gained wide acceptance internationally. Upon evaluation of these design procedures, a significant difference was found in the values of the seismic demand and the seismic capacity of unanchored liquid storage tanks. The intermediate steps in the simplified design procedures of both the API standard and the New Zealand recommendations differ significantly in approach and values. The need for an accurate method for the computation of the seismic forces exerted on unanchored tanks is apparent. Furthermore, the uplift mechanism and the method for computing both demand and allowable buckling stresses need major refinements.