Abstract:

A multi-parameter space reduced order model (ROM) of nonlinear unsteady aerodynamics is proposed based on the linear convolution and correction factor, to obtain the unsteady aerodynamics accurately and efficiently. Firstly, a linear reduced order model is constructed using the convolution of modal step responses. Then, the nearly-orthogonal Latin hypercubes are used to determine the appropriate sampling points in the parameter space, and calculate the correction factor at each sampling point. Finally, the Kriging surfaces for the correction factor are calculated using the Kriging interpolation method, and the modified reduced order model is obtained. The basic geometric model used for the validation of ROM is the Isogai wing. The effects of the Mach number, sinusoidal input amplitudes and oscillation frequency on the accuracy of the ROM are all investigated. The ROM is coupled with the structural model to make a time-marching simulation, and predict the flutter speed. The validation of the ROM in multi-parameter space is proved by the comparison with results obtained via the high-fidelity computation fluid dynamics solver.

Key words: Correction factor, Latin hypercube sampling, Kriging interpolation, Nonlinear unsteady aerodynamics, Reduced order model (ROM)