Subsystem identification in the TiltRotor from a structural numerical model provided by the Topic Manager (TM) has been the first main task. The model needed to be heavily manipulated for vibro-acoustical purposes, hence it has been refined, substructured and boundary superelements have been created as additional stiffness and mass elements.
The clustering-based subsystem identification algorithm proposed is based on the exponentiation of the Direct transfer (TD) matrix. TD matrices were calculated from the numerical model through eigenvalue decomposition. This data has been clustered by means of distance indicators and dendogram plots.
A sensitivity analysis on clustering results has been done, assessing the influence of the power order of the TD matrix and the maximum number of clusters allowed in the grouping.
In the light of results obtained, it has been concluded that the initial proposal based on exponentiation of the TD matrix has shown to be flawed for the current system, with a complete loss of the matrix structure as the power increases due to a strong coupling between all the elements defining the model. As a result, SEA requirement of weakly connected subsystems is not satisfied. However, this circumstance is often not the case in experimental scenarios. In the context of this investigation the exponentiation process can be best categorized as a method for identifying the strength of connection between already identified subsystems.
ATPA test plan has been designed: static test was already performed at the TM facilities, by instrumenting with accelerometers cabin interior panels and structural linking points of the tilt-rotor wings to the fuselage.
In parallel, local ANC algorithms have been adapted and evaluated for synthetic signals and real noise signals of a similar aircraft, provided by the TM. In this sense, it was experimentally tested a multichannel virtual sensing ANC scheme. The preliminary results were promising and will be integrated in the electronics of the ANC system.
Finally, specific electronics have been developed to make a processing with an analogical filtering to obtain an efficient, robust and low cost local ANC processing. The starting idea was to consider that the noise spectrum can vary according to different conditions: speed of the aircraft, internal equipment of the cabin, different layout of the cabin (number of seats, pilot/passenger cabin separation, acoustic treatment of the wall with absorbing materials, etc.). Results are promising.
Work is also being done to find the best way to integrate the loudspeaker and transducers in the TM seat.
