Waveform Evaluation of Seismic Network in the Aceh Province by using Power Spectral Density and Probability Density Function

Abstract

Waveform evaluation of seismic networks is important to know the station's performance. The monitoring must be carefully checked and routinely evaluated to capture some seismic activities from active faults and an unknown active fault. We develop and evaluate the seismic network in Aceh Province from three stations, namely LHMI, MLSI, and KCSI stations. We applied machine learning approaches to classify the seismic waveforms and features in the time-frequency domain and figure the power spectral density and probability density function (PSDPDF) in the time-frequency domain. The results, the characteristics of the PSD are normal and the sensor condition is in good condition. The recorded signal was well analyzed in the 19 times segmentations and figure no significant noise the vertical component. High frequency (> 10 Hz) has a low noise percentage and low signals < 0.1 Hz. A parallel trend can be found at the low frequency content and an overlapping trend at the very high frequency content. That was usually comes from ordinary equipment interference in the centering and thermal conditions. The overlapping is very meaningless because it was only located in one segment. In general, the recorded signal is suitable with the amplitude tolerance of the pole-zero seismometer configuration, so the determination of the body-wave phase signal (P and S) in the higher frequency is effective.