Experimental methodology

The sidewall of the Navigli channel was instrumented at 2 m from the base of the channel. Steel plates at 0.5 m spacing were attached to the wall with epoxy resin. Measurements were obtained from 161 points (total length L= 80 m). The wall was divided into 11 sections: ten sections of 7.5 m and the last section of 5 m long. The length of the sections was limited by the number of channels of the data acquisition system (15 channels). The accelerometers were glued to the steel plates to conduct different multichannel surface wave tests (MASW). For each MASW test, 15 traces were recorded; each trace contained 2400 data points. After testing one section, the accelerometers were moved to the next section. A one-pound hammer was used as a source to make an impact at three locations, i. e., right, center and left of the array of accelerometers. Depending on the attenuation of the wall, the source was located at least at two different horizontal distances from the first receiver.

The equipment used in all the tests (Figure 2) consisted of a digital oscilloscope (HP 35610A), four filter-amplifiers (Krohn-Hite 3984), impulse hammer (Dyatran), 15 accelerometers (PCB and Dytran, frequency range 1 Hz to 5 kHz), and a laptop-based data acquisition system of 16 bit resolutions and 1 MHz maximum sampling rate (Iotech, Wavebook E16). The sampling rate used was 62.5 kHz per channel; thus, a maximum frequency of 6.25 kHz can be recognized from the time signals and a maximum frequency of 31 kHz can be identified from the Fourier spectra. The time window was 38 ms; the frequency resolution in the spectra was 26 Hz. The time responses are 2D-Fourier transformed to determine spatial and temporal frequencies from the contour plots in the frequency-wavenumber space.