GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.1 GNGTS 2023 X=960 m with 1 m inter-receiver spacing. Here, we only consider the vertical component to compute the attributes. The data are recorded with 0.5 ms sampling interval for a duration of 0.5 s. In Fig. 1c, we show a snapshot of the vertical displacement velocity after 270.5 ms from shooting. A significant amount of energy is trapped within the fault area. In Fig. 1d, we show one of the simulated shot gathers (shot location: X=40 m). The data are significantly attenuated after passing through the fault (~offset 460 m). Results In Fig. 2, we show the estimated attributes on the tunnel floor and ceiling arrays, zoomed on the centre of the line (X= 400 m–600 m), close to the fault location. The estimated energy attribute for the floor and the ceiling are shown in Fig. 2a, whereas the autospectrum attribute, energy decay coefficient and attenuation attributes are plotted separately for the floor and ceiling arrays (Fig 2b to 2g). All attributes, show high amplitude in the proximity of the fault location and values close to zero elsewhere. The energy attribute provides a sharp increase at the fault location along each array, as expected for anomalies having lower velocity than surrounding materials (Fig. 2a). The autospectrum coherently shows two high energy spikes within the fault zone for frequencies higher than 65 Hz (Fig 2b and 2c). Similarly, the energy decay exponents (Fig 2d and 2e) show significant changes in - γ value. We plot the values of −γ such that the maxima correspond to energy concentrations and the minima correspond to energy decays. Positive and negative offset curves correspond to spatial windows located at the right and left sides respectively of each shot positions. Along both arrays, energy concentrations can be observed passing from the intact rock to the fault zone, whereas energy decays are depicted in the opposite direction. The stacked attenuation coefficients (Fig 2f and 2g) suggest strong attenuation within a significant frequency range. The frequency band affected by the anomaly is maximum at the location of the fault crossing the array, while it decreases on opposite sides for the floor and ceiling arrays. Comparing the floor and ceiling attributes, there is a slight but systematic shift in the position of the attribute anomalies. This shift is linked to the dip of the fault and to the fact that it intersects the tunnel at different locations on the floor and ceiling (Fig. 1a and points A and B in Fig. 3). We use the shift in the attribute plots and the height of the tunnel ( and in Fig. 3) to estimate the fault dip. Since the shift between the floor and ceiling attributes is sharper and easy to identify in the energy plot (Fig. 2a), we pick the position of the energy attribute maxima, which are at positions 496 m (ceiling, B in Fig. 3) and 504 m (floor, A in Fig. 3). Then, we use the distance of 8 m between the peaks to estimate the fault dip as: . (3) θ = ( ) Given that the height of the tunnel is 8 m, a (Fig. 3) of 45° is obtained from equation 3, which θ exactly matches the true angle of the simulated fault.