A fast procedure for the identification of complex mode shapes in time domain
Article
Tansel, IN, Satar, GA. (1998). A fast procedure for the identification of complex mode shapes in time domain
. 18(3), 251-258. 10.1080/02286203.1998.11760385
Tansel, IN, Satar, GA. (1998). A fast procedure for the identification of complex mode shapes in time domain
. 18(3), 251-258. 10.1080/02286203.1998.11760385
The Recursive Multichannel Maximum Entropy Method (RMMEM) has been used to identify various systems such as the cutting process, structural dynamics, and unified transfer functions of machining operations. Computational speed, small program size, and reasonable accuracy (less than 10% frequency response estimation error) make the algorithm a good candidate for microcomputer applications. This paper proposes a new procedure to estimate the natural frequency, damping ratio, and complex mode shape. The procedure has the following four steps: 1. It excites the system with a pseudo-random input signal. 2. It estimates an autoregressive vector (ARV) model by using RMMEM. 3. It obtains the digital transfer function from the ARV model and improves its accuracy by using a nonlinear least squares (NLS) algorithm. 4. It calculates the modal parameters from the discrete transfer function. Based on the cases studied, the proposed three-step procedure was found computationally efficient and accurate enough for many modelling applications. The proposed pseudo-random input signal allows very precise adjustment of the frequency content of the excitation by considering the dynamics of the system without requiring any expensive filters nor equalizers.
