Dealing with harmonics and noise for advanced dynamic identification using in-flight helicopter data

In the framework of the HC-AG19 Garteur Action Group, novel methodologies and procedures are proposed and analyzed to improve finite element models of helicopters through in-flight dynamic data. Thus, the development of methods allowing accurate modal parameter estimation is essential to provide reliable reference data to the model updating process. Recently, a complete GVT test was performed by NLR, University of Bristol and Agusta Westland on a commercial helicopter, including shaker excitation with both random and pseudo-random signals, as well as acceleration profiles recorded during specific flight maneuvers. These provide an ideal data-set for method validation and development, in particular to verify the efficiency of Operational Modal Analysis versus standard Experimental Modal Analysis techniques. An additional challenge of the data set is that the recorded in-flight data contain both the harmonic contribution coming from the rotor as well as the noise which inevitably contaminate the data when measuring in operating conditions. In this paper, harmonic removal techniques and advanced identification methods will be applied on the data. The modal models will be firstly identified using Experimental Modal Analysis to obtain a reference dataset for validation. Operational Modal Analysis is then performed on the same data but neglecting the input force to validate the methods. Different harmonic removal techniques are also applied and their efficient in removing these components from the signals compared. Finally, the simulated in-flight GVT tests were repeated using as input recorded accelerations from different maneuvers (constant speed flight, hovering, landing and rotor stop). The results in some of these conditions will also be discussed