Research on baseline drift using least-square and EMD

A major problem encountered in analog or digital accelerograms is distortion and shift of the reference baseline. Accumulated experience indicates that direct integration of the ground acceleration data provided for seismic soil-structure interaction analysis often causes unrealistic drifts in the derived displacement, which may have a significant effect on large-scale interaction analysis. A simple approach was proposed to remove long-period noise for acquiring a realistic displacement-time series. In the approach, the acceleration data were baseline-corrected in the time domain using the least-square curve fitting technique, and then empirical mode decomposition (EMD) was used to extract signal trend to remove effect produced by long-period signal. Both simulated and field measured signals were employed to demonstrate the feasibility of the approach.