ABSTRACT
Wind loading is an essential aspect in the robust design of long-span bridges, but is often not well-known and cannot be measured directly. On the other hand, the wind-induced responses can easily be measured with accelerometers in structural health monitoring systems. This paper presents the application of a newly developed Kalman filter-based algorithm for inverse estimation of wind loads and structural response. The Hardanger Bridge, a 1310 m long suspension bridge instrumented with a monitoring system for wind and vibrations, is used as a case study. The load and response behavior is represented by an augmented system model consisting of two sub-models: a reduced-order modal model of the bridge, and a state-space latent force model that characterizes the temporal evolution of the wind loading. First, the local wind field at the bridge location is analyzed in terms of spectral densities and coherence of the turbulence. The latent force model, which is built from Matérn function kernels, is fitted to approximate the spectral density of the wind loads, which can be predicted by classical theory for buffeting forces on bridge decks. Structural responses in the form of accelerations are then used in a Kalman filter to estimate the modal response and modal forces for a 30-minute data set. It is found that the estimation process is stable for acceleration output only, avoiding the accumulation of errors.
