BWIM FOR ORDINARY ROAD TRAFFIC ON A CONTINUOUS STEEL BOX GIRDER BRIDGE USING WIRELESS ACCELEROMETERS

Abstract

 Bridge Weigh-In-Motion (BWIM) enables live load assessment through bridge response measurement typically with strain gauges. BWIM based on wireless accelerometers, which can be easily installed, potentially makes the live load assessment economical and practical. In this research, BWIM which consists of automated vehicle detection and weight estimation based on only accelerometers is proposed and validated through field measurement on a continuous two-span steel box girder bridge with ordinary road traffic. Using a Kalman filter, bridge deflection is estimated only from acceleration data; the deflection is further analyzed to estimate the weight of vehicles even when there are multiple vehicles on the bridge. Then weight estimation method taking into account the speed variation is shown to have higher accuracy through numerical analysis and field measurement with test vehicles. Next, using girder-end acceleration responses, vehicle entrance to and exit from the bridge are automatically detected. The method using neural networks is shown to have higher accuracy than a threshold-based method. Finally, using both the vehicle detection and weight estimation methods, BWIM is performed on a continuously measured bridge response data. The proposed methods resulted in vehicle weights estimation performance similar to that of strain based method. By applying this method to 11 days traffic data, the traffic characteristic were clarified, which indicates the applicability of the proposed methods on ordinary road traffic.