| Abstract: | Complicated geological structures, different from the plane-parallel stratification, often occur in the underground and present problems for the geoelectrical prospecting. At least from a theoretical point of view, modern exploration procedures for dealing with this type of situations may be planned. These procedures are based on a large number of data distributed on the area to be explored, and on rigorous inversion methods, carried out by means of powerful computers.Such a procedure leads to models of the subsoil, on grounds of the apparent resistivity values measured on the surface. This is an inverse problem which implies the use, in opposite direction, of a direct computing method, i.e. a procedure allowing the computation of apparent resistivity values from a structural model. One of these, the “surface charge” algorithm is described in this paper.In practice, implementation of such ideal procedures often encounters difficulties of technical and also of economic type, at least for the 3-D problems. The consequence is that, in many cases, one is obliged to use traditional procedures, as, for instance, the use of vertical soundings distributed on the area to be surveyed.The application of traditional procedures to complicated problems requires some changes. Firstly, the popular Schlumberger array must be avoided, as it may result into insufficient information. Moreover, the array is not suitable for deep explorations because of the excessive cable length. More convenient is the dipole-dipole array, which, however, presents its own drawbacks, as, for instance, a higher sensitivity towards lateral variations. This last incovenience may be avoided by means of the transformation of the field dipole-dipole into half-Schlumberger diagrams; but it has been shown that, over arbitrary structures, this is possible only if “polar dipole-dipole” arrays are used in the field.In the final part of the paper some field examples of “continuous” polar dipole-dipole soundings carried out for deep exploration are shown. |
