Evaluation of Surrogate-Based Modeling Methods for the Optimization of Helicopter Rotor Structures for Minimum Vibration

2019-11-02
Helicopters are notorious for their high vibration levels and the rotors are the main contributors. At preliminary stages, it is essential to design the rotors to achieve minimum vibration amplitudes, which is generally realized by using optimization routines. The optimization of a helicopter rotor for minimum vibrations requires repeated high-fidelity solutions, which lead to high computational times. Moreover, since the rotor optimization problem contains many local minima by its nature, the optimization process might be repeated in order to guarantee the global minimum, which results in increased solution times. In this study, surrogate-based models for the prediction of optimization results are investigated to reduce the computational expense for the optimization of a four-blade rotor. The composite rotor blade cross-sectional design parameters are utilized for the optimization and the rotor is assumed to be in a high-speed forward flight. Along with vibration minimization, the minimum blade mass is also targeted and both of the objectives are subjected to a number of static and dynamic constraints. The objectives and the constraints are written as functions of design variables and for each function, a surrogate model is constructed. To reduce prediction errors in the surrogate model, data transformation techniques are employed. Using these surrogate models, the rotor is optimized. It is concluded that by using surrogate-based modeling and data transformation techniques, computational time required by the optimization process can be significantly reduced without compromising the optimization accuracy.
Citation Formats
M. E. Bilen, E. Ciğeroğlu, and H. N. Özgüven, “Evaluation of Surrogate-Based Modeling Methods for the Optimization of Helicopter Rotor Structures for Minimum Vibration,” presented at the 8th Asian/Australian Rotorcraft Forum, 31 Ekim - 02 Kasım 2019, Ankara, Türkiye, 2019, Accessed: 00, 2021. [Online]. Available: http://www.arf2019.org/.