New Aerodynamic Data Dispersion Method with Application to Launch Vehicle Design

This paper describes a new generalizedmethod for implementing aerodynamic data dispersions in the framework ofMonteCarloflight simulations. As opposed to the traditional pure-bias type dispersionmethods, the newproposed model is a general mathematical approach based on truncated Fourier series that, when combined with physical modeling tailored to the aerodynamic quantity of interest, enables the generation ofmore realistically dispersed data with magnitude, phase, slope variations, and a controlled amount of bias. The new method is also presented in a particular example, as applied to the Ares I-X Flight-Test Vehicle and the Ares I Crew Launch Vehicle rolling moment data. It is shown how the adoption and implementation of this newmethodwithin these projects has resulted in significant increases in predicted roll control authority and has lowered the induced risks forflight-test operations. A direct impact on launch vehicles is a reduced size for auxiliary control systems and the possibility of an increased payload. This technique has the potential of being applied to problems inmultiple areaswhere nominal data together with uncertainties are used to produce simulations using Monte Carlo type random sampling methods. It is shown that physics-based dispersion models, together with nominal data and uncertainties, can make flight simulations more realistic and allow for leaner spacecraft designs.