Bridge deflection estimation using kalman filter for data fusion

Abstract

Measuring bridge deflection under dynamic loading is useful to understand structural behavior as well as to assess health condition. However, the conventional approaches for the deflection measurements in large-scale bridges are challenging as they typically require fixed reference points that may be often unavailable in the fields. Alternatively, indirect estimation methods have been developed; other measurements such as acceleration or strain are first obtained and subsequently transformed to displacement. As the acceleration and strain are conveniently measured, the indirect estimation methods are considered to be effective and practical while they have in general lower accuracy in estimation compared to the conventional direct measurement. This study proposes a multisensor fusion approach combining acceleration and strain for bridge deflection estimation with improved accuracy. A mathematical formulation based on Kalman filtering is presented for the fusion of the different physical quantities to estimation vertical deflections. Both numerical and experimental validations are presented to show the efficacy and accuracy of the proposed approach. © 2015 Taylor & Francis Group, London.