ONLINE RESPONSE ESTIMATION USING VIBRATION DATA FROM A LIMITED NUMBER OF SENSORS FOR A COMPOSITE RAILWAY BRIDGE

Abstract eng:
The continuous monitoring of loads applied to structures and the system response is very important for structural health monitoring. For civil engineering structures, the number of sensors placed during operation is in most cases limited due to practical and economical considerations. In addition, some locations in the structure are not accessible for the installation of sensors. Combining available vibration measurements with the information obtained from a dynamic model of the structure, however, allows to estimate the response at unmeasured locations. This information can be used for the prediction of stresses and fatigue loading, as well as many other engineering applications. Recently, a joint input-state estimation algorithm was proposed by Gillijns and De Moor in [1]. The algorithm distinguishes itself from existing methods for state estimation by not making any assumptions on the unknown excitation. It was applied for response estimation in structural dynamics by Lourens et al. in [2]. In this contribution, the joint input-state estimation algorithm is applied using in situ dynamic measurements from the Sesia viaduct, a composite seven span railway bridge located on the Italian highspeed line between Torino and Milano. Local strains in the bridge cross section under the excitation of an Italian ETR500Y high-speed train are estimated, using measured vibration data and a detailed finite element model [3]. The model only includes the bridge structure. Both the dynamic loads applied by the train to the bridge and train-bridge interaction are accounted for by means of equivalent forces that are assumed for the application of the joint input-state estimation algorithm. It is shown that, by properly choosing these equivalent forces, accurate response estimates are obtained.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket