• Ottimizzazione
• Inversioni
• Forward modeling
• Dati geoelettrici
• Elettrodi sepolti

One of the main objectives of this thesis is to analyze the functionality of the Matlab code package of IP4DI, an open source software used for inversion of geoelectric data.
After a careful analysis of the implementation of both the forward and the inverse problem, it was decided to perform some changes in order to improve the calculation method for a more accurate result and also, we attempted to introduce useful potentialities not present in the software via iterative methods.
Since an important point in the inversion process is the construction of the forward modeling mesh, it was decided to pay particular attention to it. In this regard, the way in which the forward modeling grid is constructed has been changed, so that the maximum number of layers that make it up is calculated based on the maximum spacing between the electrodes and on the number of electrodes, which has been supplemented with an additional number of layers. This calculation method is introduced in order to best reproduce the conditions of propagation of the injected current, modeling the infinite boundaries.
To validate this methodological approach, we conduct various inversions starting from analytical models, the test results show that the addition of more layers refines the result of the inversion, limiting as much as possible the effects produced by the edges.
The IP4DI software allows the inversion of data both with surface electrodes and with Cross Borehole electrodes. A further approach was to implement into the code the possibility of doing the modeling prosses with buried electrodes with horizontal spreading. This modification was carried out to ensure that a more accurate replication of the real situation of the underground could be obtained in the case in which the data acquisition is performed with a horizontal spreading in a trench. Since the correct positioning of the electrodes is not considered as predominant factor, and the geometric coefficient of the surface electrodes is used as relevant instead of the one correlated with buried electrodes so that a deamplification of the resistivity values corresponding to the most superficial layers can be observed.
As a conclusion, to validate the new implementation, tests were carried out starting from analytical models. From the various tests, we have seen that the use of a software that takes into account the burial of the electrodes can be useful in particular cases, such as a model with a thin and conductive first layer. For the models in question the results of the inversions carried out both with buried electrodes and those with surface electrodes after the correction of the data reporting them on the surface, thanks to the modification of the geometric coefficient, are very similar. Therefore, in cases that are not particularly complex, the data can be corrected reporting the electrodes on the surface and than inverting the date with software that use surface electrodes.