学术文献库

A methodology of adaptive time series analysis, based on Empirical Mode Decomposition (EMD), and on its time varying version tvf-EMD has been applied to strain data from the gravitational wave interferometer (IFO) Virgo in order to characterise scattered light noise affecting the sensitivity of the IFO in the detection frequency band. Data taken both during hardware injections, when a part of the IFO is put in oscillation for detector characterisation purposes, and during periods of science mode, when the IFO is fully locked and data are used for the detection of gravitational waves, were analysed. The adaptive nature of the EMD and tvf-EMD algorithms allows them to deal with nonlinear non-stationary data and hence they are particularly suited to characterise scattered light noise which is an intrinsically nonlinear and non-stationary noise. Obtained results show that tvf-EMD algorithm allows to obtain more precise results compared to the EMD algorithm, yielding higher cross-correlation values with the auxiliary channels that are the culprits of scattered light noise.