Target-based algorithm for baseline correction of inconsistent vibration signals

Abstract

A target-based baseline correction algorithm that can directly assign acceleration/velocity/displacement values equal or close to preset target values is proposed in this study for correcting vibration acceleration signals with inconsistent initial velocities and displacements. Baseline shift or drift phenomena can arise for velocity and displacement values obtained by numerical integration of recorded acceleration signals using assumed initial values that are inconsistent with reality. Two indicators, the drift ratio and amplitude ratio, are also proposed to identify the degree of baseline drift and to evaluate the results of baseline correction. The traditional polynomial detrending algorithm has typically been used to remove unreasonable trends in time series, but insufficient attention has been paid to consistency between acceleration, velocity, and displacement, which is explicitly considered in the proposed algorithm. The target-based algorithm has two implementation schemes, that is, a precise scheme and a detrending scheme, which can be selected according to the degree of baseline drift. Three inconsistent vibration signals are considered to verify the validity of the proposed algorithm. The baseline shift or drift trends can be removed effectively using the proposed target-based algorithm with the drift ratio and amplitude ratio reduced to satisfactory ranges whereas the traditional method fails to correct these signals.