H. Bolton Seed’s research while affiliated with University of California, Berkeley and other places

The limitations of methods of earthquake-resistant design of embankments based on pseudo-static analysis and incorporating a static seismic force designated by a seismic coefficient are discussed. It is suggested that both from considerations of soil behavior under cyclic loading conditions, and for purposes of assessing embankment deformations due to earthquakes, the development of a method of predicting dynamic seismic forces and their variations with time would provide a more reasonable approach for design purposes. A method of accomplishing this objective is suggested, and representation of the results by an equivalent simplified seismic force series is proposed. For embankments that can be considered to have a uniform shear modulus, charts are presented for determination of the equivalent seismic force series. It is pointed out that limitations in available dynamic response analyses require that some judgment be exercised in selecting design parameters. However, the results presented in the paper provide a useful basis for guiding this judgment. Furthermore, the results show that in designating seismic coefficients for design purposes, it is important to distinguish between embankments of different heights and material characteristics as well as different positions of the potential slide mass within the embankment.

I am most grateful to my colleagues in the ASCE Geotechnical Division for their invitation to prepare this state-of-the-art paper on soil liquefaction during earthquakes. It is a great compliment to many colleagues and friends who through their original ideas, cooperation, and invaluable critiques have helped to raise the state of knowledge in this field over the past decade to a condition where engineers can practice in an extremely difficult area with some reasonable level of confidence.

The phenomenon of earthquake fault rupture propagation through soil is quite complex and is not well understood at this time. This paper presents the results of an integrated investigation of this problem. Insights are developed from the examination of surface fault rupture field case histories, laboratory physical model tests, and physical analogies to the earthquake fault rupture process. Field observations and experimental results illustrate the typical patterns of behavior developed in the soil overlying a base rock fault displacement. Evidence suggests that differential movement across the distinct fault rupture dissipates as the fault rupture propagates toward the ground surface through unconsolidated earth materials, and that the characteristics of the soil overlying the bedrock fault strongly influence the observed earthquake fault rupture propagation behavior.

An improved understanding of earthquake fault rupture propagation through saturated clay would assist engineers in siting and designing facilities to be constructed in regions where cohesive soils overlie potentially active faults. The results from numerical analyses suggest that the finite-element method can be applied to this class of problem provided that the soil's nonlinear stress-strain behavior is adequately modeled. It was found that the height of the shear rupture zone in the overlying saturated clay soil at a specified base rock fault displacement depends primarily on the soil's failure strain. As the clay's failure strain decreases, the shear rupture zone in the clay overlying the bedrock fault propagates further at a specified base displacement. Other material parameters such as soil shear strength and stiffness also affect the fault rupture process, but not to the extent of failure strain. The orientation of the shear rupture zone through the soil depends largely on the orientation of the underlying bedrock fault plane.

A simplified method of analysis for determining the permanent displacement of a clay core in a heterogeneous earth dam subjected to dynamic excitation is discussed. Centrifuge studies on three identical models were used to determine the deformation characteristics of earth dams subjected to earthquakes large enough to produce flow slides. These studies suggest that the failure mode is considerably influenced by the reduction in shear strength caused by the void-ratio increase and the subsequent decrease in effective stress at the sand-clay interface. Conventional sliding block analysis does not consider this phenomenon. A simplified method using the concepts of critical-state soil mechanics is presented that may be used to estimate both the reduced shear strength and the redistributed void-ratio at the interface. Values predicted using this method were found to agree well with those observed in the centrifuge experiments. By integrating these quantities into Newmark's sliding block analysis, a better estimate of the permanent displacement of an earth dam may be obtained.

This paper describes the use of 1 g small-scale models composed of weak, saturated clay mixtures to investigate the response of natural clay deposits to underlying base rock fault displacements. A 3:1 mixture of kaolinite and bentonite at water contents between 150 and 100% produces a material with undrained shear strengths on the order of 1 to 2.5 kPa and well-scaled stress deformation behavior for small-scale model testing without the need for a centrifuge apparatus. Strength testing indicated that the axial strain at failure could be controlled to evaluate the importance of this parameter in base deformation testing. It was found that the height of the shear zone in the saturated clay overlying the base rock fault offset was related to the magnitude of base movement and the failure strain of the soil.

A new simple shear apparatus is described which can apply bi-directional simple shear cyclic and monotonic loading. Cylindrical specimens are placed in a pressure chamber so that confining pressure and back pressure can be applied. Bi-directional shear loads are applied to the specimen's base, which is mounted on a series of two horizontal rolling tables which can be independently loaded in two orthogonal directions while preventing cap and base platen “rocking.” A computer-automated feedback-loop-controlled pneumatic servovalve system provides excellent control of loads and displacements. Displacement control during cyclic tests on softening specimens is improved by the inclusion of a parallel stiffness. Procedures and test results are presented for three different undrained cyclic simple shear loading conditions and for a strain-controlled monotonic loading condition.