Evangelista, Lorenza (2009) A Critical Review of the MASW Technique for Site Investigation in Geotechnical Enigeering. [Tesi di dottorato] (Unpublished)
Visibile a [TBR] Repository staff only
|Item Type:||Tesi di dottorato|
|Uncontrolled Keywords:||Surface Waves|
|Date Deposited:||24 May 2010 08:21|
|Last Modified:||30 Apr 2014 19:39|
The dispersion curves of surface waves have been successfully used for the characterization of the shallow subsurface for decades. Three steps are involved in utilizing dispersion curves of surface waves for imaging geological profiles: 1. implemented the experimental procedure, 2. create efficient and accurate algorithms organized in a basic data processing sequence designed to extract surface wave dispersion curves from accelerometric records, and 3. develop stable and efficient multimodal inversion algorithms to obtain shear wave velocity profiles. This dissertation focuses on the third step, the inversion of the dispersion curves of surface waves, with the aim of searching the best procedure to get a more accurate and reliable estimate of the geological material properties. The inversion actually is comprised of two sub-steps: 3a) estimate a model employing the theory of surface wave propagation and mathematical optimization; 3b) appraise the model for its accuracy, either deterministically or statistically. One of the major goals of this study is to find the shallow S-wave velocity structure that explains the observed dispersion curves of surface waves. This is achieved by a multimodal inversion that involves the minimization of the cost/objective function that characterizes the differences between observed and calculated dispersion data. Due to discrete nature of inversion problems, the model obtained from the inversion of the data is therefore not necessarily equal to the true model that one seeks. This implies that for realistic problems inversion really consists of model estimation followed by model appraisal. General speaking, there are three catalogs of inversion techniques based on the internal physical principle of the geotechnical problems: linear inversions, non-linear inversions, and trial and error methods. It is common to present the inverted Vs profile as a unique profile without showing a range of possible solutions or some type of error bars, such as the standard deviations of the Vs values of each layer. Additionally, the person performing the inversion usually assumes the prior information required to constrain the problem based on his or her own judgment. Implementing an inversion method that includes estimates of the standard deviations of the Vs profile and finding tools to choose the prior information objectively were the main purposes of this research. To perform SASW inversion, one global and one local search procedures were presented and employed with synthetic data: a pure Monte Carlo method. The synthetic data was produced with a forward algorithm used during inversion. This implies that all uncertainties are caused by the nature of the MASW inversion problem alone since there are no uncertainties added by experimental errors in data collection, analysis of the data to create the dispersion curve, layered model to represent a real 3-D soil stratification, or wave propagation theory. The pure Monte Carlo method was chosen to study the non-uniqueness of the problem by looking at a range of acceptable solutions (i.e., Vs profiles) obtained with as few constraints as possible. It is important to note that this method requires large amounts of time to obtain Vs profiles with low rms error. Based on the variety of shapes found for Vs profiles with satisfactory rms, the non-uniqueness of SASW inversion was evident, concluding that the dispersion curve does not constrain the solution sufficiently to determine a unique Vs profile or to resolve specific velocity contrasts between layers. A summary of the reasons for this factor : 1. Characteristics related to the experimental dispersion curve: 2. Number and distribution of data points describing the experimental dispersion curve 3. Uncertainties of the experimental dispersion data 4. Characteristics related to the initial shear wave velocity profile 5. Depths and thicknesses of the layers 6. Depth to half-space 7. Initial shear wave velocities The points that represent the dispersion characteristics of a site needs to be selected carefully to have: (i) sufficient data to include all important features of the dispersion curve, and (ii) a good balance of information content to resolve the Vs of the layers based on similar amounts of information and have a fairly weighted rms error that gives a good measure of the fit between theoretical and experimental data. Therefore, using a multimodal inversion algorithms can reduce the non-uniqueness of the SASW inversion, allowing to use all the information contained in the experimental dispersion curve. To improve the interpretation step of the MASW experimental procedures, it is focused the attention of the dispersion behaviour of Rayleigh wave in complex startigraphic condition. It is analysed the influence of subsoil structure on the experimental data and the error introduced in the inversion step, assuming a one-dimensional soil stratification. Two dimensional numerical models are developed to investigate the behaviour of Rayleigh waves in the presence of lateral anomalies: slope interface between the layers. Different geometrical configuration are analyzed to take account also the influence of the source, relating to the immersion of the layers. The results show a clear influence of the subsurface structure, that can induce to a underestimation of the correct thickness layer of the soil profile investigated. To overcome this limitation a new approach is proposed to correct the dispersion curve with adequate factor before the inversion problem.
Actions (login required)