Vibration monitoring during the excavation by Drill & Blast of a tunnel close to an urban centre: analysis of the results

According to previous research, the exponential k statistical law could approximate much better than classical site laws the behaviour of blast induced vibrations during the excavation of a tunnel and therefore predict the PPV, but not enough data were available to demonstrate this statement in the previous research. Thanks to geotechnical surveys performed by GD Test s.r.l. Company, after almost 10 years from the previous campaign, a consistent number of recorded data during the realization of a tunnel below the city of Merano (province of Bolzano) was available to be analysed, allowing to make further considerations. All the analyses done were performed using a Monte Carlo method generating 10.000 random combinations of 4 parameters of k statistical law to determine each time the best set which better describes the experimental data used. Recorded data were divided mainly according to three lithological zones with RMR varying suddenly in the same zone and in some meters. Ranges of confidence found were different, however they have always shown a minimum of 70% of confidence level which have increased till over 95% reducing the area analysed, especially in intervals below 100 m, and grouping data based on both the RMR and lithological dominium. Additionally, to the demonstration that k statistical law could well describe experimental data with some limitations, it was demonstrated that RMR is strictly correlated with the data to obtain good confidence level, indeed three ranges of RMR for each lithological dominium were defined and then for each range were selected only data recorded by geophones approximately positioned in places that hypothetically have the same RMR. This last test has proven that zones with same RMR not only at the tunnel face blasted but also where geophones are positioned can be well described by the k statistical law that therefore could be a good predictor equation for zones with same RMR. From this analysis it's possible to observe that a limit of the k statistical law is the use at the same time of many data for which the level of confidence drops down below 70%. Therefore, as demonstrated thanks to many data it was possible to demonstrate that k statistical could be a good site law with some limitations, but also, as a novelty, that it could represent very well the prediction equation for zones with homogeneous RMR.