Abstract
This research proposes a structural identification method with variable geometric parameters extending the concept of self-identification.
The method is formulated by using the change of eigenvalues and eigenvectors according to the change of variable geometric parameters.
The method is then numerically examined through the identification of the stiffness matrix of a two-dimensional variable geometry truss.
As a result, the stiffness matrix is identified with a set of linearly independent eigenvectors corresponding to the one vibration mode changed by one variable length member of the truss.
Also, the results show that the identification performance using only one variable length member is relatively equal to that using three variable stiffness members without geometric change when multiple vibration modes data are available.
Further, it is shown that the identification error with the variable length member depends on the vibration mode and the stiffness ratio between the stiffness of the variable length member and that of the other members to be identified.
Finally, the increment of degree-of-freedom of the identified matrix to the identification error is investigated.
