Key points are not available for this paper at this time.
A method is presented for identifying parameters in a linear, discrete of a structure by using measured normal modes to modify an analytically derived model. The structure considered has a relatively large number of points of interest and a frequency range of interest influenced by a relatively small number of normal modes. The analytical which is introduced has fewer degrees of freedom (normal modes) than coordinates (points of interest). The parameters of this incomplete model are obtainable from the limited, but quantitative, test data, and the conceptually valid, but approximate, analytical data. The characteristics of this and methods of using it are discussed, in particular, means of computing the effects of mass and stiffness changes on natural frequencies and normal modes. Computer experiments illustrate these methods and demonstrate that such a can be useful and that the procedure is not overly sensitive to measurement errors.
Berman et al. (Sun,) studied this question.