Noise Floor Characterization in Accelerometers for Earthquake Monitoring

Earthquake prediction and monitoring are increasingly needed for humans because of their fatal effects. Given the unpredictable nature of these phenomena, early detection of possible damaging seismic waves is crucial and, therefore, has been an increasingly main focus of research. Timely detection of earthquakes comes through seismic stations in addition to algorithms. These stations contain extremely expensive instrumentation, which limits their widespread use. Nowadays, however, many sensors have been designed to be capable of making the same measurements as seismographs at a significantly lower price. In order to verify the applicability of MEMS accelerometers in structural building monitoring and earthquake prediction, it is mandatory to characterize their noise floor. Given the extensive comparison of the strengths and weaknesses of the two accelerometers already discussed by many authors, this paper will compare their performances for earthquake detection applications. A noise floor characterization of these sensors is presented. The objective is to perform a prior metrological evaluation of these instruments, comparing low-cost MEMS-based and analog accelerometers used to date. A comparison of noise floors using FFT and Allan Variance will be reported. The good results show these instruments' applicability in the field, paving the way for their more widespread distribution.