GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 3.2 99 by the presence of artifacts in the images, as well as by the deviation from the true resistivity of the subsoil. The resolution of an acquiring sequence of measures is strongly influenced by the sensitivity distribution, which can be very different. In recent years many authors dealt with the problem of optimization of the data sequences of acquisition, which can be formed by various combinations of measurements with different electrode arrays and/or different dipolar lengths and distances, in order to obtain the best resolution and at the same time the best recovery capacity of the tomographic section. It emerged, therefore, the need of the study of the sequences of measurement, for the choice of optimized dataset, aiming at a higher resolution of investigation and, especially, at a more efficient retrieval of information on the characteristics of the medium investigated. Recently many studies have been carried out to compare the advantages and limitations of the most common arrays used in electrical tomography. Park and Van (1991) and Van et al. (1991) have pointed out the difficulty of acquiring unnoisy field data using the pole-pole array. Sasaki (1992) compared the resolution of pole-pole, pole-dipole and dipole-dipole arrays in cross-hole electrical tomography, ruling that the dipole-dipole array is more suitable for the resolution of complex structures. Comparing the same arrays, Oldenburg and Li (1999) have pointed out the differences in the depth of penetration of each array, considering the inverse models obtained. Studies on the resolving power and reliability of Wenner array (Dahlin and Loke, 1998; Olayinka andYaramanci, 2000) emphasized the importance of the sampling density to determine the resolution of this configuration. Over the past two decades, developments inmicroelectronics have led tomajor improvements for both acquisition systems and inversion software. Consequently, the tomographic resistivity techniques have achieved remarkable results. Some recent multichannel resistivity-meters enable many simultaneous measurements for each current dipole, significantly reducing the measuring time and, at the same time, allowing the acquisition of big datasets to cover large areas with high sampling density. These multichannel systems consequently has allowed the use of new multielettrode arrays and sequences, Examples of these types of array are the moving gradient and the midpoint-potential-referred. Dahlin and Zhou (2004) have carried out several numerical simulations to compare the resolution and efficiency of pole-pole, pole- dipole, Wenner, Schlumberger, dipole-dipole, moving gradient and midpoint-potential-referred array. They recommend moving gradient, pole-dipole, dipole-dipole and Schlumberger array, rather than the other, although the final choice should be determined by the type of geology expected, the purpose of the survey and other logistical considerations. Martorana et al. (2009) have compared various 2D simulations using near surface models, by varying the array type, in order to assess the best resolution. Despite the flexible nature of these systems, there still is the tendency to acquire apparent resistivity measures using traditional electrode arrays such Wenner, Schlumberger or dipole- dipole. These latter often prove to be a good choice, as their properties are well known, in terms of depth of investigation (Barker, 1989), lateral and vertical resolution (Barker, 1979) and signal-to-noise ratio (Dahlin and Zhou, 2004). However, the choice of these arrays could not be the most efficient if the time or the number of steps permitted for the survey is limited, or if a target of particular interest is spatially localized. The use of parameters to estimate the reliability of inverse models is crucial to the choice of the measures sequence which provides reliable results, while not entailing excessive costs, that are linked essentially to the number of current dipoles used and survey time. Recently, there has been a growing interest in the study of the generation of sets of measures for optimizing the resolution tomographic image. One of the benefits of the optimization of the sequential experimental design would be a better reconstruction of the subsurface that aims to reduce the number of acquisitions of the data and, consequently, the cost of the survey, without compromising the quality of the reconstruction itself. In practice the methods of array optimization to maximize the resolution of tomographic image should also take into account the