Filtering techniques for extracting bridge frequencies from a test vehicle moving over the bridge

Abstract

By letting a test vehicle move over a bridge, the frequencies of the bridge can be identified by picking the peaks in the Fourier response spectra of the test vehicle. One problem with this approach is that the vehicle frequency (undesired) usually appears as a dominant peak in the spectrum, making it difficult to identify the bridge frequencies (desired). To enhance the visibility of bridge frequencies, an effective filter is needed to filter out the vehicle frequency. As a preliminary attempt in this study, three filters are adopted, i.e., the band-pass filter (BPF), singular spectrum analysis (SSA), and the singular spectrum analysis with band-pass filter (SSA-BPF), proposed herein as a combination of the above two. Through the numerical study on two cases with the vehicle frequency smaller or larger than the 1st bridge frequency, the SSA-BPF technique is demonstrated to be most effective for extracting the bridge frequencies, due to its salient feature that there is no need to select the number of singular values, as required by SSA, while the unexpected peaks encountered by the BPF are avoided.