Eigensensitivitity analysis of large-scale structures by substructuring method

In sensitivity-based model updating, the eigensolutions and the associated derivatives of the analytical model need to be calculated iteratively to achieve an optimal solution, which is very time-consuming for a large-scale structure. To overcome this shortcoming, this paper proposes a new method to calculate the eigensensitivity via substructuring approach. When the design parameters in one particular substructure need to update, the sensitivity matrices to the updating parameters at this substructure are calculated while they are zeros at other substructures. The eigensensitivity at the global level is assembled by performing some constraints on the sensitivity matrices at the substructure level. The proposed method is applied in a frame and a practical bridge structure. Numerical results show that the present substructuring approach is more efficient than the traditional global method, at the satisfactory precision.