Bridge displacement estimation by fusing accelerometer and strain gauge measurements

• Published in: Structural control and health monitoring Vol. 28; no. 6
• Main Authors: Ma, Zhanxiong, Chung, Junyeon, Liu, Peipei, Sohn, Hoon
• Format: Journal Article
• Language: English
• Published: Pavia Wiley Subscription Services, Inc 01.06.2021
• Subjects: Acceleration measurement; accelerometer; Accelerometers; Algorithms; Beam theory (structures); data fusion; Data integration; displacement estimation; Displacement measurement; Euler-Bernoulli beams; Field tests; finite impulse response filter; FIR filters; Frequency drift; Parameter estimation; strain gauge; Strain gauges; Transformations
• ISSN: 1545-2255; 1545-2263
• DOI: 10.1002/stc.2733

Summary For large‐span bridge monitoring, displacement measurement is essential. However, it remains challenging to accurately estimate bridge displacement. When displacement is calculated by the double integration of acceleration, a low‐frequency drift appears in the estimated displacement. Displacement can also be estimated from strains based on the Euler–Bernoulli beam theory. However, prior knowledge of the mode shapes and the neutral axis location of the target bridge are required for strain–displacement transformation. In this study, we propose a finite impulse response filter‐based displacement estimation technique by fusing strain and acceleration measurements. First, the relationship between displacement and strain is established, and the parameter associated with this strain–displacement transformation is estimated from strain and acceleration measurements using a recursive least squares algorithm. Next, the low‐frequency displacement estimated from the strain measurements and the high‐frequency displacement obtained from an acceleration measurement are combined for high‐fidelity displacement estimation. The feasibility of the proposed technique was examined via a series of numerical simulations, a lab‐scale experiment, and a field test. The uniqueness of this study lies in the fact that the displacement and the unknown parameter in strain–displacement transformation are estimated simultaneously and the accuracy of displacement estimation is improved in comparison with those of previous data fusion techniques.