Date of Award

12-2014

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Civil Engineering

Advisor

Dr. Ronald D. Andrus

Committee Member

Dr. C. Hsein Juang

Committee Member

Dr. Nadarajah Ravichandran

Committee Member

Dr. WeiChiang Pang

Abstract

Published results of laboratory tests to evaluate the influence of the at-rest lateral stress coefficient (K0) and the overconsoildation ratio (OCR) on liquefaction resistance of sands are reviewed in this dissertation. It is found that significant increase in liquefaction resistance can occur with high K0 and/or high OCR. A unique relationship between liquefaction resistance of isotropic and anisotropic normally consolidated sands exists based on the published laboratory data. Relationships between liquefaction resistance of overconsolidated sands and normally consolidated sands also have been shown in previous laboratory studies. Based on the previous studies, expressions for K0 and OCR correction factors (KK0 and KOCR, respectively) are recommended. Methods of estimating in situ K0 in sands are reviewed. An improved relationship for estimating K0 is derived based on results of dilatometer tests (DMT), cone penetration tests (CPT) and OCR using published calibration chamber test data. Values of K0 estimated from DMT-CPT-OCR in natural and man-made sand deposits are compiled and compared with values of K0 calculated from self-boring pressuremeter test (SBPMT) results. Estimates of K0 based on both DMT-CPT-OCR and SBPMT increased at about 4 to 5% per log cycle of time based on the geologic age of the sand deposits. Methods of estimating K0 based on the ratio of horizontally to vertically polarized shear wave velocities propagating in the horizontal directions (VsHH/VsHV) are reviewed. When plotted versus the geologic age or the time since the last liquefaction event, values of VsHH/VsHV do not exhibit an increasing trend, but a slight decreasing trend. These results suggest that both VsHH and VsHV increase at similar rates with time, and thus the influence of age is cancelled out in the ratio. An improved relationship for estimating K0 from VsHH/VsHV is suggested by adding the age term and assuming average values of the fabric anisotropic ratio and the stress exponent. Because K0 increases with age, it is reasoned that much of the K0 correction on liquefaction resistance of aged sands is captured by the age or diagenesis correction factor KDR that has been proposed by various researchers. Hence, it is recommended that the KK0 factor not be used with the KDR factor, unless there is clear evidence that KK0 is not captured by KDR. The entire effect of OCR is not captured by KDR, however. Therefore, it is suggested that the KOCR factor be used in field liquefaction evaluations of aged sands. Values of K0 are estimated from DMT-CPT-OCR information in natural sand deposits at eight geotechnical investigation sites in the South Carolina Coastal Plain (SCCP). Estimates of K0 for the eight sites fall in the typical range for natural sands. A gentle increasing trend in K0 with age is observed in these data, similar to the previously published data. A weaker trend is observed in K0 with the time since initial deposition or the time since the last liquefaction event, which indicates that K0 decreases little, if any, because of liquefaction. Based on the VS profiles from the SCCP sites, the crosshole VsHV (velocity of vertically polarized shear waves propagating in the horizontal direction) and the SCPT VsVH (velocity of horizontally polarized shear waves propagating in near vertical direction) measurements are in good general agreement with each other. Ratios of VsHH to VsHV at six crosshole sites indicate greater overall stress stiffness in the vertical direction than in the horizontal direction. Values of K0 estimated from ratios of VsHH to VsHV at the SCCP sites fall in the typical range for natural sands when the age term is added to the predictive relationship. Profiles of compression wave velocities (VP) suggest existence of unsaturated (i.e., VP < 1,400 m/s) zones extending from 0.2 to 1.3 m below the groundwater tables at five of six crosshole sites in the SCCP. The liquefaction potential of sand deposits in the SCCP is evaluated based on both penetration resistances and VS, assuming the ground shaking levels during the 1886 Charleston earthquake. Five of the eight sites are no liquefaction sites. The other three are paleoliquefaction sites, with only one site exhibiting surface manifestations of liquefaction in 1886. Geological ages of the sand deposits at the eight sites range from 70,000 to over 1,000,000 years. Overconsolidation ratios range from 1.0 to 2.4 in the critical sand layers. The behavior at six of the eight sites during the 1886 Charleston earthquake is correctly predicted when corrections for diagenesis, unsaturated conditions and OCR are made. One of the no liquefaction sites and one of the paleoliquefaction sites are incorrectly predicted to liquefy in 1886. It is possible that liquefaction, or at least pore pressure build up occurred at these two sites in 1886, but the capping non-liquefiable layer was thick enough to prevent noticeable surface manifestations. The results support the use of diagenesis, unsaturation, and OCR corrections, as well as the need to assess the capping layer and the critical layer thicknesses in liquefaction assessment of aged sand deposits.

Share

COinS