Finite element model updating of a truss model using incomplete modal data

This paper presents an application of a two-step finite element model updating methodology to a 3D truss bridge model, with the inertial moment and elastic modulus as variables. First, the sensitivity of the change in modal parameters to variations in the physical parameters were analyzed. Based on these sensitivity analyses, the parameters that can be updated were determined. Next, the optimization of an orthogonallybased objective functional was employed in model updating using the modal data obtained from measurements on a physical structure. Sequential Quadratic Programming (SQP), a quasi-Newton approach that creates a linear search strategy applicable to non-linear minimization problems was used to acquire a precise and stable solutions. In computing the solution, the gradients of every parameter versus the objective function were determined by executing a quadratic fitting methods at each iterative step in order to improve the solution’s precision. The results of the experimental updating show that the analytical model is improved after model updating such that the natural frequencies from the finite element model better match the experimentally determined natural frequencies.