Inversion algorithm Sample Clauses

Inversion algorithm. For the sites where both high-quality Rayleigh wave dispersion and H/V ratio curves were derived, a joint inversion scheme, as that proposed by Parolai et al. 2005 and 2006, was used to estimate the local S-wave velocity profile. In particular, the non-linear inversions were performed using a genetic algorithm, which does not rely upon an explicit starting model and allows the identification of the solution close to the global minimum. Here, the genetic algorithm proposed by ▇▇▇▇▇▇▇▇ and ▇▇▇▇▇▇ (1996) was used. The forward modelling of Rayleigh wave phase velocities and H/V curves was performed using the modified Thomson-▇▇▇▇▇▇▇ method proposed by ▇▇▇▇ (1999) and following the suggestions of Tokimatsu et al. (1992) and ▇▇▇▇ and ▇▇▇▇▇▇▇▇▇ (2004), under the assumption of vertically heterogeneous 1D earth models. The validity of this assumption is investigated by computing the H/V curve for each station of the array. Whenever the H/V curves provide the same estimate of the fundamental frequency and the main peak exhibits a good level of shape similarity, it is reasonable to assume that the geological structure underneath the array does not change significantly. Thus, in these cases, the resulting average H/V curve can be safely considered representative for the medium’s volume underneath the array. The modelling of both the dispersion and H/V ratio curves during the inversions was not restricted to the fundamental mode only, but the possibility that higher modes can participate to define the observed dispersion curve is allowed. During the inversion, the two data sets are weighted in a balanced way using the cost function proposed by ▇▇▇▇▇▇▇▇ et al. (1999). Obviously, when one of the two experimental curves cannot be used (e.g. because of the presence of man-made industrial noise signals) the proposed inversion scheme can still be applied using only one of either the dispersion or H/V curves. In our cases, dispersion curve and H/V curves were both used at all sites with the exception of LGN and BZZ. At these two sites only dispersion curves were considered. For further explanation on surface wave methodologies, see document: Project S4: ITALIAN STRONG MOTION DATA BASE, Deliverable # 6, Application of Surface wave methods for seismic site characterization, May 2009. a) Example of installation of an EDL acquisition systems plus Mark-L4-C-3D sensor. b) Set-up of the stations. c) and d) Example of array geometry in Faenza. d) shows also stations which worked prop...