Proceedings of the 27 th Seminar on machinery vibration, Canadian Machinery Vibration Association CMVA 09, Vancouver, CB, 16 p. 1 SOFTWARE FOR OPERATIONAL MODAL ANALYSIS AND AUTOMATIC IDENTIFICATION OF MODAL PARAMETERS V.H. Vu 1 , M. Thomas 1 , A.A. Lakis 2 and L. Marcouiller 3 1 Department of Mechanical Engineering, École de technologie supérieure, Montreal, QC H3C 1K3, Canada 2 Department of Mechanical Engineering, École Polytechnique, Montreal, QC H3C 3A7, Canada 3 Hydro-Québec’s Research Institute, Varennes, QC J3X 1S1, Canada ABSTRACT In this paper, we present a software for the Operational Modal Analysis (OMA) of vibrating structures in operating conditions. The method used is based on a multivariate autoregressive model, with the model’s parameters of the model are estimated by least squares via the computation of the QR factorization, and the modal parameters are identified from the eigen- decomposition of the state matrix. The natural frequencies, damping rates and modes shapes are updated with respect to the model order and are successively constructed on stabilization diagrams with their corresponding confidence intervals. Furthermore, an optimal model order can be automatically selected from the evolution of a factor called the Noise rate Order Factor (NOF) from which the structural modes are automatically distinguished from the spurious ones in order to construct noise-free spectra. After the frequency ranges of interest are selected, the natural frequencies and damping rates are automatically identified. The proposed software is user friendly and the operator can easily determine the accuracy of the modal parameters that are automatically computed. Several experimental applications are described by way of examples. Corresponding author: Email: marc.thomas@etsmtl.ca 1. INTRODUCTION The identification of structural modal parameters [1] plays an important role in structural health monitoring and machinery vibrations. It is usually conducted in a frequency domain by measuring the transfer function between a vibratory response and a known excitation force [2]. However, in several industrial applications where it is not possible to stop a machine, the forces cannot be measured, and so an Operating Modal Analysis (OMA) must therefore be conducted [3, 4]. Since the environmental forces result from natural excitations, operational modal analysis should deal with the time domain. Several methods, such as the Ibrahim time domain method (ITD) [5], the Least squares complex exponential (LSCE) [6] and ARMA [7], can be used in identifying of modal parameters just from responses. Examples of such industrial applications