The Application of the SVD-FDD Hybrid Method to Bridge Mode Shape Estimation

Abstract

AbstractThis study conducts a numerical simulation and a field experiment to measure the bridge’s traffic vibrations and to estimate the mode shapes of the bridge by the FDD (Frequency Domain Decomposition) method. But traffic vibration does not always satisfy the assumption to apply the FDD method, the results were not expected to be calculated correctly. So we use the SVD (Singular Value Decomposition) as the support to improve the applicability of the FDD method for traffic induced vibration. In the numerical simulation, the modal-parameters were estimated by acting traffic loads to a finite element plate model using shell elements. As expected, the mode shapes and natural frequencies estimated by the FDD method only were not accurate. However, we were able to improve the accuracy by comparing those results with the estimation results of the SVD method. In the field experiment, a vehicle and a bridge with known specifications were used to conduct multipoint bridge vibration measurements at 10 points using MEMS (Micro-Electro-Mechanical Systems) sensors. The results of the field experiment showed a similar trend to the numerical experiments, we could most likely estimate the natural frequencies of the bridge.KeywordsTraffic-induced vibrationSVDFDDModal parametersNatural frequencyMode shapeMAC