FE Model Updating of Cable-Stayed Bridges Based on the Experimental Estimate of Cable Forces and Modal Parameters

The paper presents an example of model updating of both the mass and stiffness parameters of a curved cable-stayed bridge in Venice (Italy). Conventional optimization problems of mass and stiffness using ambient vibration data are prone to ill-posedness and ill-conditioning, and generally, the scholar must assume one of the two to achieve a reliable estimate. However, it is possible to assess the mass and stiffness from ambient vibration tests in cable-stayed bridges following a two-step procedure. In the first step, the scholar can assess the mass matrix from the cable forces estimated from the natural frequencies of the cables. Then, the unscaled mode shapes and natural frequencies are used to tune the stiffness matrix in a second step. The authors proved this updating approach with two variance-based sensitivity analyses. The former is the sensitivity of the cable forces to the specific mass and bearing deformability. The latter is the sensitivity of the natural frequencies to Young’s moduli. Two global optimization algorithms for mutual validation, differential evolution (DE) and particle swarm optimization (PSO), are then implemented for the model calibration.