Parametric estimation of systems with modal interference

The topic of this paper is the accuracy improvements in parametric estimation of systems with modal interference. A channel-expansion technique has been previously used in the Ibrahim time domain method to well perform the modal identification. However, the error involved in the determination of system order from the frequency response function or the Fourier spectrum associated with each of the response channels would generally lead to a distortion in the modal identification, especially for a system with modal interference. In the present paper, the singular value decomposition in conjunction with least-squares analysis is introduced in the procedure of Ibrahim time domain method to obtain the system matrix and determine the system order. Also, the phase angle diagram of frequency response function can be employed to distinguish the structural modes influenced by modal interference and then avoid the phenomenon of omitted modes from the distortion of system order determined by frequency response function. Numerical simulations, including two examples of a model of the motor vehicle, and a linear two-dimensional model of one-half of a railway vehicle, confirm the validity of the proposed method for modal identification of a system with modal interference.