In-flight Modal Identification by Operational Modal Analysis

Operational modal analysis (OMA) has been widely used in many fields of study because it allows identifying the modal parameters of a flexible structure in its operating condition. The system is under unknown working loads assumed to be random with broadband spectral characteristics. These hypotheses are not always easy to fulfill, generating uncertainty about identified modal parameters. This study evaluates and compares the effectiveness of two OMA techniques, enhanced frequency-domain decomposition (EFDD) and Ibrahim time domain (ITD), in the accuracy of modal parameter estimation of an unmanned aerial vehicle (UAV) structure with output-only data obtained by flight testing. To evaluate the influence of the number of sensors used in the identification of the modes, different measurements setups were considered to carry out in-flight modal identification analyses. Some works have addressed uncertainty by focusing on retesting or subdivision of a single measurement record. This work innovates in presenting an uncertainty study considering the variables that intervene in the estimation of PSD. The uncertainty in the identified modal parameters is obtained using the variability of the values of the parameters found. The modal frequencies values observed employing EFDD and ITD do not present substantial variations associated with the PSD matrix estimates. The EFDD damping ratio values show significant variability because they are mainly affected by spectral leakage, while the ITD damping ratio values are less sensitive to Welch's method parameters variation. The root mean square deviations (RMSDs) of the frequencies values for both techniques are compared with those resulting from ground vibration testing.