The available simplified analytical methods for the seismic design of underground 5 structures either assume yielding or rigid-unyielding conditions. Underground reservoir 6 structures do not fall into either of these categories. In this paper, we present the results of three 7 centrifuge experiments that investigate the seismic response of stiff-unyielding buried structures 8 in medium dense, dry sand and the influence of structure stiffness and earthquake motion 9properties on their performance. The structure to far-field spectral ratios were observed to 10 amplify with increased structural flexibility and decreased soil confining pressure at the 11 predominant frequency of the base motion. Lateral earth pressures and racking displacements for 12 a range of structural stiffnesses were compared with procedures commonly used in design. Pre-13 earthquake measured lateral earth pressures compared well with expected at-rest pressures. 14 However, none of the commonly used procedures adequately captured the structural loading and 15 deformations across the range of stiffness and ground motions for which these reservoirs must be 16 designed. Further, it is unclear if the current methods of analysis provide conservative or 17 unconservative results for engineering design purposes. This identifies a critical need for 18 improved methodologies to analyze and design underground reservoir structures. ManuscriptClick here to download Manuscript Journal paper manuscript_vf.docx 4 of these structures are not fully captured by simplified seismic design procedures. Soil-structure-5 interaction (SSI) for these buried structures is complex and depends on foundation fixity, 6 properties of the surrounding soil, flexibility of the structure relative to soil, and the 7 characteristics of the earthquake motion. There is an increasing need in engineering practice to 8 obtain a better understanding of the seismic performance of these underground structures. For 9 example, the Los Angeles Department of Water and Power (LADWP) is replacing some of its 10 open water reservoirs with buried, reinforced-concrete reservoirs to meet water quality 11 regulations. Understanding the seismic performance of these restrained underground structures 12 will improve the structural and geotechnical seismic design of these type of projects. 13 Traditionally, underground structures are categorized either as yielding or rigid-unyielding, 14 and are designed differently based on the categorization. A yielding wall is one that displaces 15 sufficiently to develop an active earth pressure state. The current state of practice for assessing 16 seismic earth pressures on yielding structures relies heavily on the Mononobe-Okabe (Okabe 17 1926; Mononobe and Matsua 1929) and Seed-Whitman (Seed and Whitman 1970) methods. For 18 rigid-unyielding walls that don't undergo any deformation, the method of choice is often the 19 simplified solution proposed by Wood (1973), which assumes a completely rigid wall (with no 20 flexure). Underground reservoir structures fall in ...

The difficulties associated with instrumenting earthquake sites in order to record pore pressure changes in a future event led to the use of scaled model tests performed in a centrifuge. Both dry and saturated sands were employed, contained in a box constructed of aluminium laminae designed to move freely on each other. This would result in shearing distortions developing in the soil unimpeded by the container. Accelerometers, displacement transducers and pore pressure sensors were attached to the box and embedded in the soil at various elevations so as to record the response of the soil to an earthquake-like excitation supplied to the base of the container. A special apparatus was constructed to imitate earthquake motion. In some tests on saturated sand, the soil profile was liquefied. Test results of accelerations, lateral and vertical displacements and pore pressures against time for typical earthquake inputs are given. The data, obtained under controlled conditions, can be compared with the various calculation methods for dynamically generated pore pressures.

SUMMARYExperimental and analytical studies were conducted to determine dynamic soil-structure interaction characteristics of a single-span, prestressedconcrete bridge with monolithic abutments supported by spread footings. The experimental programme, consisting of harmonic forced vibration excitation of the bridge in the transverse and longitudinal directions, revealed the presence of four modes in the frequency band, 0 to 11 Hz, and the onset of a fifth mode at 14 Hz, the highest frequency attained during the tests. The fundamental mode at 4.7 Hz was the primary longitudinal bending mode of the deck and had a relatively low damping ratio that was approximately 0 0 2 5 of critical. The second and third modes at 6.4 Hz and 8.2 Hz were the primary twisting modes of the deck which involved substantial transverse rocking, transverse translation and torsion of the footings. As expected, the damping ratios associated with these two modes, lz = 0.035 and l3 = 0.15, were directly related to the relative amounts of deck and footing motion. The fourth mode at 10.6 Hz was the second twisting mode of the deck and involved relatively little motion of the footings and abutment walls, which was consistent with the low damping, c4 = 0 0 2 , observed in this mode. The response data at 14 Hz suggested that the fifth mode beyond this frequency was the second longitudinal bending mode of the deck involving longitudinal translation and bending of the abutment walls.A three-dimensional finite element model of the bridge, with Winkler springs attached to the footings and abutment walls to represent the soil-structure interaction, was able to reproduce the experimental data (natural frequencies, mode shapes and bridge response) reasonably well. Although the stiffnesses assigned to the Winkler springs were based largely on the application of a form of Rayleigh's principle to the experimental data, these stiffnesses were similar to theoretical foundation stiffnesses of the same size footings on a linearly elastic half space and theoretical lateral stiffnesses of a rigid retaining wall against a linearly elastic backfill.

No abstract

No abstract