Modified Ibrahim Time Domain Method for Identification of Closely Spaced Modes: Experimental Results

Abstract

Traditional Experimental Modal Analysis (EMA) is generally known as a standard tool to identify the dynamic signature of structures. Frequency domain methods are most based on Fourier analysis which transforms the time data to the frequency data. The methods such as: Ibrahim time domain (ITD) method, the least-squares complex exponential (LSCE) method and the Eigen realization algorithm (ERA), are the well-known methods in time domain that require free decay responses or Impulse Response Function (IRF) of the structures. It is often claimed that ITD does not work well in case of closely spaced modes because the method is SIMO, whereas the other mentioned techniques that are known to be MIMO are claimed to work well also in case of closely spaced modes. In this paper, a modified ITD is proposed using Hankel matrix averaging for identification of closely spaced modes. The traditional ITD method is investigated under the presence of closely spaced modes and is compared to a modified formulation of the ITD that can take advantage of multiple input loadings. It is shown that the results of the traditional ITD has been improved using the modified ITD in identification of closely spaced modes. To validate the proposed method, an experimental case study with closely spaced modes in a gear box shaft is considered.