Operational Modal Parameter Estimation from Short-Time Data Series

The field of Operational Modal Analysis (OMA) has recently become an emerging research interest. OMA, also known as Response-Only Modal Analysis, extracts the modal parameters by processing only the system response data. In many experimental measurement situations, the system’s output data series is short in length and buried under potentially correlated noise. In such cases, the separation of the noise from true data is challenging, generally resulting in inconsistent modal parameter estimates. For many of these measurement situations, it is believed that the correlated noise is a function of the system's operation and that a linear Auto-Regressive with eXogenous input (ARX) model exists that should describe the true system output. In this paper, a Nonlinear Auto-Regressive with eXogenous input (NARX) model based approach for estimating modal parameters from short output time series data is explored. In this approach, nonlinear terms are added to a linear ARX model to describe the noise and essentially filter out the system’s true output data from the noisy time data series.