Development of seismic design criteria for category I cofferdams

Abstract

Gravity-type cellular cofferdams are widely used as waterfront retaining structures. Static design of these structures is generally performed by considering several modes of static failure. The experience gained from post-construction observation of cofferdam behavior has greatly enhanced the postulation of admissible failure mechanisms that should be considered in design. In nuclear power plant operation, assuring the safety and stability of waterfront embankments is important in providing an undisturbed supply of cooling water under all seismic and flood conditions. In moderate to highly seismic areas, the protection of the waterfront with an economic class I embankment may be quite difficult or impractical as the area required to develop the relatively flat slope construction that is required for seismic stability is often unavailable. Therefore, more attention should be directed toward the design and construction of cellular cofferdams as waterfront earth retaining structures. However, in the absence of data on dynamic behavior of cofferdams under seismic conditions, the methods for design and construction of structures which meet the current regulatory requirements are not well documented. Because static methods of cofferdam design have become somewhat standardized and have been verified through observations of performance, a promising approach to the seismic design of such structures is centered about the development of consistent pseudo-static failure mechanisms. This paper illustrates such extensions of the static design techniques, and compares the factors of safety obtained from static and pseudo-static (seismic) analyses. Certain design criteria, which have been modified to account for dynamic action, are explained. Local soil, rock and hydrostatic conditions, liquefaction and hydrodynamic forces are considered. Design/analysis parameters are suggested for evaluating extreme condition seismic loading, i.e., the Safe Shutdown Earthquake (SSE). Included among the parameters are; active and passive dynamic earth pressure coefficients; location of groundwater and free water surfaces to be used in conjunction with the SSE; the coefficient of friction acting at the interface between the fill material within the cell and the material on which the cell is founded; dynamic pressure distributions due to groundwater and free-standing water adjacent to the structure; vertical and horizontal coefficients of seismic acceleration. In addition, the consequences of postulated liquefaction of adjacent materials are investigated, and measures are suggested to adjust the analysis to accommodate such an occurrence. Among the postulated failure conditions which are considered under seismic loading are the following: • sliding, • overturning, • slippage between the sheeting and the cell fill, • shear failure along the centerline of the cell, • Cummings method of horizontal shear, and • interlock strength. The method of analysis admits various assumptions that are based on the static failure modes of the analytical models; for example, the postulated sliding failure condition assumes an inflexible rigid body type of behavior while the Cummings method of horizontal shear considers the structure to be flexible and the soil to be capable of developing failure planes. By extending these postulated failure modes for the dynamic analysis, the same modeling assumptions are admitted for the pseudo-static case. The methodology proposed is suggested as a means of conservatively performing design checks, using simplified procedures, and not as a replacement for a dynamic, state-of-stress response analysis.