Permanent displacement estimation method based on variational mode decomposition

In the near-field area of strong earthquakes, the information recorded by a strong seismometer includes ground motion information and complex low-frequency noise. This noise results in the presence of an “extra” acceleration component in the record that is independent of ground motion, and these components are integrated into the velocity and displacement time history to cause offset. In order to solve this problem, this paper adopts a baseline correction method based on variational mode decomposition (VMD), which decomposes seismic signals into modal functions of different frequencies and then extracts the low-frequency modal functions. It is worth noting that these low-frequency modal function components contain linear trend terms, a property that allows us to perform baseline corrections efficiently. The results show that the permanent displacement obtained by this method is in good agreement with the coseismic displacement observed by GPS stations nearby. This method provides an efficient approach for estimation permanent displacement which helps revealing the focal mechanism of an earthquake.