GNGTS 2018 - 37° Convegno Nazionale

706 GNGTS 2018 S essione 3.2 The test using firework charges was also performed on the top of the rock face using the same channels and the same sampling frequency, but the shots were triggered in 8 different positions near geophone 1 and geophone 2. Charges were fired in natural surface cavities or fractures. Filtering was applied as in the previous test, except for the notch filter, since no harmonic noise was disturbing these records. Tab. 2 summarizes the results for all the employed channels. Comparing the results of the two tests, it is obvious that the hammer has a better performance than firework charges. Using the hammer, except a few channels for Hammer 1 and Hammer 2 hits, all the 5 geophones record the vibrating signals, while for the trigger test with firework charges, all the 5 geophones could record the signals only in Explosion 3. Therefore, by now hammer is the preferred source for the tomographic survey. Tab. 2 -Performance of each channel to sense the signal for different shots performed with firework charges. Trace no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Channel Trig 1X 1Y 1Z 2X 2Y 2Z 3X 3Y 3Z 4X 4Y 4Z 5X 5Y 5Z Explosion1 Y Y Y Y N N N N N N N N N N N N Explosion2 Y Y Y Y N N N N N N N N N N N N Explosion3 Y Y Y Y Y Y Y Y Y Y N Y N Y Y N Explosion4 Y Y Y Y Y Y Y Y N N N Y Y N N N Explosion5 Y Y Y Y Y Y Y N N N N N N N N N Explosion6 Y Y Y Y Y Y N N N N N N N N N N Explosion7 Y Y Y Y Y Y Y Y Y Y N N N N N N Explosion8 Y Y Y Y Y Y N N N N N N N N N N Event location with a uniform velocity model . The seismic source localization algorithms have been reviewed by considering the most common earthquake location techniques. The probabilistic, non-linear, global-search earthquake location method, implemented in the NonLinLoc software (Lomax et al. , 2009), was used in this work. Since no velocity information was available, event location was performed to preliminarily test the accuracy of this method, and the data from the trigger test with the hammer were used. For the Probabilistic Density Function (PDF), the Equal Differential Time (EDT) likelihood function was used to estimate the event location. The standard deviation of the uncertainties for observed arrival time and calculated travel time at each observation was assumed to be 4.5 ms. We manually picked the P-wave first arrival to obtain the observed arrival time, while travel time calculation was based on the assumption that the ray path between the source and the receiver is a straight line, instead of considering the Eikonal finite-difference scheme. The simple grid search method was used for travel time calculation. The 3D grid model was derived from the Digital Terrain Model (DTM) of the rock face, having size 200×200×200 m and 100 nodes in each direction. A constant velocity is assigned to the geometric model. According to the typical values of P-wave and S-wave velocities (Hardy and Reginald, 2005), we assumed P-wave velocity in limestone to be 3200 m/s, and we set the velocity in the air to 331 m/s to take into account the area of the model above the DTM surface. The estimated source position was determined by the grid node having the maximum value of the PDF. The result of event location is shown in Fig. 2. The five hammer hits are located next to the real hit point and detailed location information can be found in Tab. 3. The coordinates of the 5 analyzed events are the same along the X- and Y-axis and show a minor variation along the Z direction of just 2 m, i.e. one grid spacing. In terms of the average error in each direction, Y direction has the minimum error value and Z direction shows the largest error, but all the 1D average error values are within one grid spacing. For the 3D error, the average value is only