Drive‐by frequencies extraction by means of synchrosqueezed wavelet transform

Abstract

Bridges and viaducts are facing growing traffic, increasing loads as well as ageing and deterioration. Nowadays, vibration‐based Structural Health Monitoring (SHM) represents an effective solution to extract quantitative information regarding bridge structural condition. Bridge frequencies are fundamental parameters capable to reflect and summarize bridge dynamics and health condition. Typically, their identification is performed by direct approaches, which rely on the use of sensors installed on the structure under analysis. A cost‐efficient, movable, and flexible alternative is offered by drive‐by methods, consisting in the use of instrumented vehicles to extract bridge frequencies. This paper presents an experimental campaign carried out on a multiple‐span mixed railroad bridge, with the aim to identify its frequencies through onboard measurements. A set of MEMS wireless accelerometers is deployed on the vehicle, firstly allowing for vehicle pitch component mitigation. Vehicle accelerations were then used to extract bridge frequencies, by means of a Most Probable Modal Frequencies approach, relying on synchrosqueezed wavelet transform. At a cruise speed of 30 km/h promising results in terms of indirect frequency estimation were found, with no traffic disruption.