Metodologías multi-offset para la mejora de resultados de georadar y desarrollo de nuevas aplicaciones

Institución otorgante:

Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales

Fecha:

2019-04-08

Tipo de documento: 

info:eu-repo/semantics/doctoralThesis
 

Formato:

application/pdf

Idioma:

spa

Temas:

GPR - ARREGLOS SINTETICOS DE EMISORES - OFFSET VARIABLE - GPR - SYNTHETIC EMITTERS ARRAY - VARIABLE OFFSET

Descripción:

The main objective of this Thesis is to carry out studies on multi-offset (MO) methodologies of reflection Georadar (GPR) to: 1) improve the signal-to-noise ratio and the continuity of the primary reflections in soils and structures that show low response levels and 2) evaluate the potential of these methodologies in relation to new GPR applications. We investigated the use of MO methodologies to study cylindrical structures with significant curvature, such as trunks and pillars. These structures produce abundant multiple reflections that, in single-offset (SO) data sections, cannot be distinguished from the primary reflections of interest and that can hide them due to their relatively high intensities. We have solved that ambiguity through the simultaneous acquisition of MO profiles, and the modeling and adjustment of theoretical curves, which allowed identifying the multiples. This adjustment, made also possible to increase the precision in the calculation of the velocity of propagation, which in turn allows better positioning the reflectors. By adapting migration and visualization procedures commonly used in prospecting of low-curvature air-ground interfaces, it was possible to generate final images of the sections of the structure. As an application of this methodology, the internal structure of the cylindrical pedestal of a monument was investigated. The objective of the study was to detect and position possible reinforcement elements along the pedestal, in a reliable and precise way. This was important to design a safe disassembly and relocation plan by the specialists. Due to the aerial and downtown location of the monument, the ambient noise predominated over the GPR signals. This noise could be eliminated from the data sections by an iterative adjustment of sinusoidal waves on the traces, and the subsequent subtraction of them. The application of the joint SO-MO methodology allowed identifying numerous primary reflections at reinforcement elements of the structure, inhomogeneities of its material, multiples, and reflections external to the structure, with high reliability and in detail. The final images of the pedestal sections showed the distribution of the reflectors inside, clearly and precisely, thus solving the continuity/discontinuity of the reinforcement elements along the pedestal. Second, we studied the joint application of a MO methodology, the one-dimensional Emitter Synthetic Arrays (SEA-1D) method and rotation properties of the scattering matrix (PRMS) to improve the calculation of the orientation of linear structures, such as pipes and poor junctions, from data acquired along a single survey line. This is important in cases where it is not possible to prospect the terrain using 2D grids, due to obstacles on the soil surface, such as fixed objects and significant uneveness of the terrain, and a greater precision in the result is required than that obtained from the preceding SO-PRMS method. In this sense, the use of the SEA method is aimed at reducing the fluctuations of the primary reflection, which is produced by the interference of clutter and noise, and then to increase the precision in the result. As an initial step, the characteristics of the SEA-1D wave fronts that are necessary for an effective improvement of the SO reflections were analyzed, in particular, their continuity and amplitude level with respect to the noise and clutter were analized, as functions of the main parameters of the arrays, that is, the number of emitters, and the distance and time shift between them. A procedure for optimizing the SEA results was defined through indicators of the improvement, which made the results independent of the interpreter's criteria. Regarding the previous SO-PRMS methodology, the SEA-1D-PRMS implementation significantly reduced the fluctuations in the resulting azimuth distributions. An implementation of the Common Midpoint and PRMS methods (CMP-PRMS) was also evaluated. The comparison of numerical and experimental results has shown that the SEA-PRMS methodology leads to azimuth values that are more accurate than those of the other methods. Finally, the SEA-1D method has been extended to 2D arrays. The objective of this extension was to combine the beneficial effects produced on the primary reflections by 1D arrays oriented along the survey line (1Dx) and transversally to it (1Dy). Array structures and relationships between the components that allow simple and efficient acquisition of data and optimization of results were analyzed. The effects of varying the most relevant parameters of the array were studied. The SEA-2D results were compared to those of the SEA-1Dx, SEA-1Dy, CMP and SO. In particular, two fundamental soil models, that includes a large reflector and a small diffractor, respectively, were analyzed. It was demostrated that the SEA-2D arrays significantly improve the results of the other methodologies, both in relation to the continuity of the signals and the signal-to-noise level.

Identificador:

https://hdl.handle.net/20.500.12110/tesis_n6681_Bullo