Force estimation in bridge substructure boundary under vehicle loading using interacting filtering approach

Abstract

Vehicle-induced vibration may cause fatigue in bridge structures leading to sudden failure causing loss of economy and human lives. Structural fatigue estimation is a complicated proposition since the entire structure needs to be monitored irrespective of the fatigue proneness of different parts. This eventually invites huge computational costs due to high-dimensional modeling and also dense instrumentation for precise estimation of the fatigue-related properties or parameters. It has been perceived that only certain specific locations are fatigue-prone and standalone fatigue monitoring of only that subdomain is possible provided the forces acting on to it are available. To isolate the subdomain of interest from the entire structural domain, one needs to compensate with boundary forces at the domain boundaries which is never possible to be measured. This study proposes a subdomain estimation-based approach powered by an interacting Ensemble-Particle filtering approach (IP-EnKF) that monitors only a subdomain of interest (fatigue-prone) independently for its structural and damping properties while estimating the boundary forces in parallel for subsequent fatigue life estimation. This allows the employment of computationally inexpensive predictor models for any model-assisted health monitoring approach while reducing the required cost and effort of instrumentation. The approach has been validated on a bridge structure excited by a 4-degrees-of-freedom half-car model, and subsequently, unknown boundary forces are estimated for the fatigue life assessment. The proposed method estimates the boundary force under vehicle–bridge interaction and has been found effective and reliable.