Student: Chiara Smerzini
Supervisors: Prof. Roberto Paolucci
The aim of this work is to illustrate an empirical procedure for evaluating transient ground strains and rotations based on the records obtained by two dense seismic networks, namely the Parkway Valley, New Zealand, and the UPSAR, California, arrays. Due to the substantial lack of directs measurements of strains and rotations it is common practice to derive them indirectly relying on simplified relationships with restricting assumptions. The dynamic behavior of the surface ground strain tensor is investigated, introducing invariant measures of peak ground strains: the highest and lowest principal strains, HPS(t) and LPS(t), respectively. The computed peak ground strains, PGSs, show an important dependence on azimuth, by a factor of about two. Furthermore, a clear linear relationship of the PGS as a function of the most common measures of ground motion severity, such as peak ground acceleration PGA, velocity PGV and displacement PGD is found whatever the parameter. The results of the interpolation procedure are in reasonable agreement with other published relationships, arrayderived estimates and with some of the few available direct records of strain in buried pipelines and tunnels. Nonetheless, such relationships point out the limitations of the simplified approaches used in the engineering practice, which uniquely consider the wave passage effect on the ground strain evaluation and tend to underestimate significantly the observations. A similar linear trend is found when PGV data are correlated with Peak Ground Rotation, PGR, estimates. Provided the scarce knowledge concerning the issue of rotational ground motions, more attention is devoted to a broad synthesis of maximum rotation estimates from different sources. Although the empirical estimates for PGR are in reasonable agreement with other array-derived estimates, synthetics obtained by 3D Spectral Element simulations and individual data retrieved from literature. Nevertheless, the comparison between synthetics and experimental data shows some crucial differences, such as the potential dependence on site effects, which should be checked against a larger number of experimental data and numerical simulations.
You may download a digital version of this MSc dissertation here.