SISMALARM 5.0 TEST REPORT

Size: px

Start display at page:

Download "SISMALARM 5.0 TEST REPORT"

Rafe Bell
5 years ago
Views:

1 SISMALARM 5.0 TEST REPORT Center for Seismological Research (CRS) National Institute of Oceanography and Experimental Geophysics (OGS), Udine, Italy - 11th June

2 Object Evidence of SismAlarm at Center for Seismological Research (CRS) of National Institute of Oceanography and Experimental Geophysics (OGS). Date and place 28/06/2018 Udine Partecipants Leonardo Ubaldi, Roberto Pozzi, Maurizio Taormina, Marcello Taormina with the assistance of Professor Giovanni Costa University of Trieste, Department of Mathematics and Earth Sciences - and Davide Zuliani - the Electronic Engineer at the "Center for Seismological Research (CRS)" of the OGS, Udine Trieste. Editor Leonardo Ubaldi Report Date 06/07/2018 Purposes The purpose of the evidence is to test the "Guardian SismAlarm 5.0" device (hereinafter called SismAlarm) on a vibrating table. The tests are targeted to define the minimum frequencies of oscillations and the sensitivity on the acceleration values detectable by the SismAlarm. These data are compared with those detected by the professional instrument with which the vibrating table is equipped; finally, data of two real earthquakes are simulated with the two instruments to verify and compare the results. Description The National Institute of Oceanography and Experimental Geophysics owns an horizontal axis vibrating table able to be programmed to reproduce specific oscillations and oscillations coming from real data of earthquakes. 1

Picture 1: the vibrating table with Sismalarm The SismAlarm is fixed on the vibrating table on which there is a professional sensor with laser technology which is useful for comparing the data.

3 Picture 1: the vibrating table with Sismalarm The SismAlarm is fixed on the vibrating table on which there is a professional sensor with laser technology which is useful for comparing the data. The tests were performed simulating sinusoidal oscillations from 10Hz to 0,1Hz with variable acceleration depending by the frequency due to the limits of the seismic table. It was analyzed the behavior of the sensor in time and in frequency. After that, two oscillations were simulated taking data from two real earthquakes, one of which was the 24th of August in L Aquila. 2

4 Oscillation at 10Hz: The results obtained on the sensor on the seismic table are: While the values obtained from the SismAlarm are: Graphic 1: Oscillation at 10Hz in T from simulator Graphic 2: Oscillation at 10Hz in T from SismAlarm Considering that in the graphic obtained from Sismalarm each unit in the time axis corresponds to 2,7ms, it is counted a period of 100ms corresponding to the upper graph, while for the amplitudes the graphic already takes into account the sensitivity and resolution of the accelerometer, and has a value of about 5 m/s 2 almost identical to the values of the simulator graph. 3

5 About the trend of the frequency, the results obtained from the simulator are: While the trend of frequency from SismAlarm is: Graphic 3:Oscillation at 10Hz in F from Simulator Graphic 4: Oscillation at 10Hz in F from SismAlarm On the axis of abscissas are indicated the Hz, while on the axis of the ordinates are indicated the values of accelerations; it is show a peak at 10 Hz with the value of 4,92 m/s 2. Both of values in the two graphics are absolutely comparable. 4

Oscillation at 1Hz: Graphic 5: Oscillation at 1Hz in T from simulator Graphic 6: Oscillation at 1Hz in T from SismAlarm By decreasing the frequency there aren t differences in the acceleration

6 Oscillation at 1Hz: Graphic 5: Oscillation at 1Hz in T from simulator Graphic 6: Oscillation at 1Hz in T from SismAlarm By decreasing the frequency there aren t differences in the acceleration detection. While it should be highlighted, that the wave form is very dirty, due to frictions presents on vibrating table. Even the SismAlarm can detect these oscillation of very small intensity. 5

7 Graphic 7: Oscillation at 1Hz in F from Simulator Graphic 8: Oscillation at 1Hz in F from SismAlarm In the two graphics it is shown the main tone at 1Hz and secondary tones at about 3Hz due to the friction of the vibrating table. The amplitude detected in the time domain correspond to that detected in the frequency domain. 6

8 Oscillation at 0,5Hz: Graphic 9: Oscillation at 0,5Hz in T from simulator Graphic 10: Oscillation at 0,5Hz in T from SismAlarm Graphic 10(sx) and 11(dx): Oscillation at 0,5Hz in F of Simulator (sx) and of SismAlarm (dx) At 0,5Hz SismAlarm can still detect the oscillation with value of acceleration of 0,3 m/s 2 as peak. 7

SismAlarm still manages to detect such slow and low intensity oscillations, even if the background noise, due to friction, is

9 Oscillation at 0,2Hz: Graphic 12: Oscillation at 0,2Hz in T from simulator Graphic 13: Oscillation at 0,2Hz in T from SismAlarm Graphic 14(sx) and 15(dx): Oscillation at 0,2Hz in F of Simulator (sx) and of SismAlarm (dx) At the frequency of 0,2Hz, the SismAlarm still manages to detect such slow and low intensity oscillations, even if the background noise, due to friction, is much more important at these low levels of intensity. Looking at the graph in the frequency domain, we notice a very defined line at 0.2Hz. 8

the trend-time it is difficult to recognize the sine-curve at 0,1Hz because the noise seems to have comparable levels.

10 Oscillation at 0,1Hz: Graphic 16: Oscillation at 0,1Hz in T from simulator Graphic 17: Oscillation at 0,1Hz in T from SismAlarm Graphic 18(sx) and 19(dx): Oscillation at 0,1Hz in F of Simulator (sx) and of SismAlarm (dx) To go down a lot in frequency to reach 0,1 Hz, in the trend-time it is difficult to recognize the sine-curve at 0,1Hz because the noise seems to have comparable levels. Otherwise, in frequency there is a peak centered at 0,1Hz but looking the graphic in detail we note that its value is only 0,014m/s 2, and it is very low. 9

the graph generated by the simulator, while in the graphic generated by the Sismalarm the noise covers part of the information.

11 Simulation of an Earthquake: Graphic 20: Earthquake in T from simulator Graphic 21: Earthquake in T from SismAlarm It will be loaded on Seismic Table data of a Real Earthquake where the Primary Wave and the Secondary Wave are well distinguishable in the graph generated by the simulator, while in the graphic generated by the Sismalarm the noise covers part of the information. We try to analyze in more detail the results obtained in 4 phases as in the image below: Graphic 22: Split of sisma in 5 phases 10

12 PHASE A: Pre-Earthquake Graphic 23: Phase A in Time Graphic 23: Phase A in Frequency No particularities are noted in the frequency response, namely there is no spectral content in the area of interest. 11

13 PHASE B: Pre-Earthquake + Primary Wave Graphic 24: Phase B in Time Graphic 25: Phase B in Frequency It is shown that between 5Hz and 13Hz is present relevant spectral content, but it is attributed to the Primary Wave. 12

14 PHASE C: Primary Wave Graphic 26: Phase C in Time Graphic 27: Phase C in Frequency The spectral analysis is completely comparable to the previous one, even in the presence of the single Wave P. 13

15 PHASE D: Primary Wave + Secondary Wave Graphic 28: Phase D in Time Graphic 29: Phase D in Frequency Analyzing both P-wave and S-wave, there is a spectral content much higher, most centered between 4Hz and 8Hz, but with intensity more elevated. 14

16 PHASE E: Secondary Wave Graphic 30: Phase E in Time Graphic 31: Phase E in Frequecy Analyzing only the S-Wave there is a peak at 4Hz and the successive at 8Hz. 15

17 Whole Wave: Graphic 32: Whole Wave of Simulator Graphic 33: Whole Wave from Sismalarm Above there are the two spectra performed over the entire period of oscillations of 4096 samples. We have noted peaks at 4Hz and 8Hz presents only in the analysis of S Wave. There is full correspondence between the analysis performed by the Simulator and the one with the Sismalarm. 16

Earthquake simulation in Amatrice: Graphic 34: Oscillation of the Amatrice s Earthquake in T from simulator Graphic 35: Oscillation of the Amatrice s Earthquake in T from Sismalarm Graphic 36:

18 Earthquake simulation in Amatrice: Graphic 34: Oscillation of the Amatrice s Earthquake in T from simulator Graphic 35: Oscillation of the Amatrice s Earthquake in T from Sismalarm Graphic 36: Oscillation of the Amatrice s Earthquake in F from simulator Graphic 37: Oscillation of the Amatrice s Earthquake in F from SismAlarm Taking as example a real earthquake, that one occurred in Amatrice on 24th of August 2016, we reproduced the data detected by the seismograph positioned on the epicenter (the P- Wave will not be seen). Even in this case the two spectral graphs are very similar. 17

19 Conclusions: The SismAlarm gave exceptional results comparable to those of professional instruments, being able to go down a lot in frequency (0.1Hz) and detecting oscillations with very low intensity (0, 04m/s 2 0,041g). Analyzing the data in the frequency domain, we can still distinguish seismic waves from disturbances as they have significantly different frequencies. By applying special data analysis strategies, we can distinguish between disturbances and anthropomorphic actions, even low intensity seismic events. 18

Response spectrum Time history Power Spectral Density, PSD

Response spectrum Time history Power Spectral Density, PSD A description is given of one way to implement an earthquake test where the test severities are specified by time histories. The test is done by using a biaxial computer aided servohydraulic test rig.