Edinburgh Napier University

  Bouckovalas et al. (2009) is the most recent design methodology for drains for liquefaction remediation and is based on the work of Seed and Booker (1977) for infinitely permeable drains and uniform soil with purely horizontal drainage. This paper briefly summarizes the main assumptions of the revised methodology and presents a modified set of design charts, after examining the effect of the empirical exponent (A) in the relationship describing the excess pore pressure built ru,g versus N/NL. Resultes from 3-D dynamic fully coupled numerical analyses of the seismic response of a thin liquefiable sand layer improved with gravel drains are presented, providing a verification background for the Bouckovalas et al. (2009) methodology. The above numerical analyses are used for the evaluation of the coefficient of volume compressibility mv,3. Namely, mv,3 is evaluated based on (i) simulation of liquefaction element tests - a.cyclic triaxial (TX) and b. (direct) simple shear (DSS), (ii) dynamic loading of the gravel drain improved sand layer followed by a consolidation analysis immediately after the occurrence of the peak excess pore pressure and (iii) back calculation of the mv,3 coefficient using the Bouckovalas et al. (2009) analytical methodology. It is found that the coefficient of volume compressibility mv,3 may be considered approximately constant in the usual design range of ru,max = 0.2 - 0.5. In addition, the TX estimated prove to overestimate the mv,3 coefficient by at least two times, whereas the DSS predictions lie closer to the actual boundary value problem predictions.