Simultaneous monitoring of displacement and rotation of a long-span bridge based on innovative integration of GNSS and accelerometer measurements

Monitoring both displacement and rotation of a large bridge structure under different loadings is important to ensure the health condition of the structure. It is, however, very challenging to accurately measure the dynamic displacement and rotation in real-time simultaneously. We propose a new method for doing this accurately by an innovative integration of GNSS (Global Navigation Satellite System) and accelerometer measurements with the Kalman filtering technique. This approach expands the state vector to include displacement, velocity, acceleration, and rotation terms. By incorporating rotation parameters, a link between the accelerometer reference frame and the GNSS reference frame is then formulated, allowing the misalignment effects to be easily corrected. Additionally, a dual-rate measurement update scheme is devised to enable the real-time estimation of both rotation and displacement simultaneously. Experimental results using datasets from a shaking table and a real-world cable-stayed bridge recorded during a heavy vehicle load test have shown that the proposed approach could mitigate effectively the misalignment errors between the sensors, the most important error source in such measurements, and measure precisely the five-degree-of-freedom structural deformations. The controlled shaking table test has shown that the displacement measurement accuracy was improved for up to 22 % and 56 %, respectively, compared to a standalone GNSS system or an accelerometer and GNSS integrated system based on the conventional data processing method. The loading test on a cable-stayed bridge has indicated that the accuracy of the rotation measurements could reach up to 0.002°.