ACT/FHS system identification including engine torque and main rotor speed using the PBSIDOPT method

Abstract

The identified models of DLR's research rotorcraft ACT/FHS have been improved constantly over the last years. Nevertheless, the current models still have deficits that are attributed to missing engine dynamics. Therefore, in this paper the influence of engine torque and main rotor speed on model fidelity and model structure is investigated by identifying two linear models of the ACT/FHS. The first model's dynamics and outputs are identified using the rigid body states, engine torque and main rotor speed. The dynamics of the second model are identified using only the rigid body states. This model includes torque and rotor speed only as additional outputs that are not weighted during the identification of the model's dynamics. To avoid the definition of a model structure beforehand, the optimized predictor-based subspace identification method is used as system identification method with dedicated flight test data of the ACT/FHS. The results of this paper are used to clarify if torque and rotor speed are necessary for high fidelity system identification using other identification methods. Furthermore, the experimental setup, the PBSIDopt method and the model selection process are described briefly.