Summary The construction of the new ’C’ Line of Rome Underground is an important design and realisation challenge, particularly for the ’T2’ and ’T3’ Lots that will run below the historical centre of the city, in view of the enormous archaeological, historical and artistic heritage existing on the surface and in the immediate subsoil. The need to safeguard such a heritage against possible adverse effects due to the realization of the new line was felt so strongly by the Authorities in charge of its conservation that they forced the General Contractor to constitute a Technical-Scientific Committee (TSC), of which I am honored to be a member, with the task of studying specifically the interactions between the new Line and the monuments. To this aim, the TSC coordinates the activities of some Work Groups made of specialists in the areas of Geology, Geotechnical and Structural Engineering, Preservation and Restoration of Artistic and Architectural Assets, and Geomatics. The logical route followed by the TSC and the Work Groups to study the interaction between the Line and the monuments and to select the most appropriate safeguard measures implies a sequence of activities that takes into account the exceptional value of the preexisting structures and the nature of the subsoil that will host the new line. Some significant monuments existing along the path of the above-mentioned Lots are shown in Figures 2 and 3. To this suggestive surface path corresponds an analogous sub-surface path, that physically hosts the new line. In a city like Rome the immediate subsoil hides a second, invisible city, heritage of a millenary past, that occasionally becomes visible thanks to archaeological excavations and must be protected against the effects of the realisation of the new line. At larger depths starts the natural subsoil environment, that hosts the tunnels of the new Line. In the T3 Lot starting from S. Giovanni, the tunnels run first inside deposits of Pleistocene and Olocene fine-grained soils, then they dip, although only partially, inside basal Pliocene clays, and finally they emerge, before the Colosseo, in the upper Pleistocene layers made of medium- to fine-grained deposits (Fig. 5). In the T2 Lot, starting from Piazza Venezia, the tunnel path runs inside the recent deposits of the Tevere valley, comprised of mainly clayey and sandy sediments (Fig. 6), characterized by poorer mechanical properties. The interaction studies have mainly been addressed to the evaluation of possible damages to existing monuments on the surface and, to this aim, reference has been made to the damage classification proposed by BURLAND [1997]. More specifically, average damage evaluations have been performed with Burland classification, that does not require a detailed modeling of the structure, whereas more specific evaluations, of local nature, have been obtained through more advanced structural analyses based on detailed geometrical models. These evaluations have been performed in sequence according to a methodological approach that is based on firstand second-level analyses. The first-level analyses, less accurate, have been carried out assuming that the deformed configuration of the buildings is the same as that of the subsidence trough under green-field conditions. This kind of assumption generally produces the most conservative estimation of the damages to the buildings, that are forced to reach the maximum values of the deflection ratios and of the tensile strains. If, under these conditions, the level of estimated damage is higher than “negligible damage”, a more accurate second-level analysis is started, in which the stiffnesses of both the soil and the building are taken into account in the interaction. Second-level analyses are carried out using geotechnical and structural models adequate to reach the desired goals. Moreover, in these analyses drainage phenomena in both the soil and the lining due to the tunnel construction are considered, together with the soil deformations that progressively arise as a consequence of these phenomena. 50 BURGHIGNOLI RIVISTA ITALIANA DI GEOTECNICA As concerns the soil, a non-linear, elasto-plastic model with isotropic hardening has been adopted. The model takes into account, among other things, of the progressive decay of the soil stiffness with the increase of the deviatoric strain and of the stiffness variation in the unloading/reloading cycles. This model, that has been extensively tested, is implemented in the commercial code Plaxis. The choice of structural models has been more articulated since, due to the need of achieving a high geometrical detail in order to ascertain also locally possible structural damages and the consequent computational burden, a preliminary estimate of the accuracy achievable with linear and non-linear models has been sought, the latter being more accurate but also definitely heavier from the computational point of view. Such a preliminary estimate has been obtained through parametric analyses, employing a structural model based on damage mechanics. It has thus been possible to define an average damage functional, that depends on the tensile strains, and to study its evolution as a function of the curvature imposed to the basis of the entire masonry wall. This functional has been calculated performing both linear and non-linear analyses of the structure, coming to the conclusion that linear models can be considered sufficiently reliable when the damage level remain rather low, typically lower than “negligible damage”. In order to employ the most appropriate geotechnical and structural codes, the interaction studies have generally been carried out identifying first a solid having a simple shape and the same footprint on the ground as the actual building and equivalent to the latter in terms of weight and stiffness (equivalent solid). Such equivalent solid has then been employed in the geotechnical analyses to study the interaction arising from the excavation of the tunnels. The displacement field of the equivalent solid has then been imposed to the foundations of the actual building to evaluate its response in the framework of a structural analysis. As an example, the most significant results obtained from the interaction analyses concerning the Basilica di Massenzio, included in the T3 Lot, and the Palazzo dell’Amministrazione Doria Pamphilj, included in the T2 lot, are reported here. The construction of the Basilica di Massenzio (Fig. 26) was started by Massenzio in 300 b.C. and then finished by Costantino. During the IV and V centuries the Basilica was subject to modifications, however in the VI century it was already abandoned. It is a very heavy building, with high contact pressures, transferred to the soil by foundations constituted by the extension of the walls. First-level analyses (Figs. 32 and 33) have lead to an evaluation of the average damage that can be classified as “negligible”, according to the classification of Burland. On the other hand, studying the local response of the monument, firstlevel structural analyses (Figs. 34 and 35) revealed particularly high stress and strain states, due to the self-weight of the building, that give rise to a high level of damage, to which the effects of the excavation of the tunnels are superimposed with a negligible increase in the level of damage. These results should not be surprising, considering the large number of partial collapses of the Basilica that occurred in the past and its still high residual structural vulnerability. Second-level analyses, based on an accurate geotechnical characterization of the subsoil (Figs. 28, 30, and 31) and on the use of the equivalent solid, have confirmed the results of the first evaluations, revealing even more critical effects, as concerns the average damage, than those resulting from first-level analyses. The second example concerns the Palazzo della Amministrazione Doria-Pamphilij (Fig. 42). This is a building constituted by four bodies realised between the XVI and the XIX centuries. The foundations start at about 7 m from the present ground level. First-level analyses led to an evaluation of the average damage level in the range between “negligible” and “medium”. These results have been substantially confirmed by similar structural analyses for the evaluation of the local damage (Fig. 46). Second-level analyses required a suitable geotechnical characterization of the subsoil (Figs. 47 and 48) and have been performed with geometrical 3D models and with the use of the equivalent solid (Fig. 49). The results of these analyses have again led to an evaluation of the damage as “negligible”, mainly localised in the foundation walls in the hogging area (Figs. 50 and 51). Particular attention has been paid to geotechnical remedial measures aimed at compensating and attenuating the expected damage. Such measures were adopted in those cases in which a higher risk of damage to the structures existing on the surface was expected and, in any case, for buildings having (or hosting objects) of exceptional historical or artistic value. Taking into account the need to act in real time, if on the basis of a continuous and accurate monitoring differences were found with respect to the expected effects, the implementation of the compensation grouting technique has been envisioned, which consists in injecting controlled volumes of mortar in the subsoil, between the tunnels and the buildings on the surface, in order to compensate the subsidence induced by the excavations (Fig. 53). To reduce within acceptable limits the displacements induced in the surrounding soil by the deep excavations needed for the realisation of the stations, the use of sacrificial panels (crosswalls) has been planned. Crosswalls, which consist of diaphragms made of mixtures of poor concrete, are placed against the peripheral diaphragms to reduce their deformations, then they are demolished during the phases of excavation within the peripheral diaphragms and of construction of the internal floors.