Determining Aircraft Sizing Parameters through Machine Learning

Abstract

Aircraft conceptual design is an inherently iterative process as many of the methods employed do not have an analytical solution. Additional design requirements increase the computational cost of these iterations, especially in the case of unusual aircraft configurations; traditional tube-and-wing designs rely heavily on the use of empirical correlations for initial design performance, while more unusual designs with less data available require the direct use of expensive physics-based tools in order to properly estimate aircraft performance. This leads to a desire to minimize the number of iterations when evaluating a new concept. In some cases, it is advantageous to split the optimization process into an optimization loop and a sizing loop, where the sizing loop contains design parameters that do not work well in the traditional optimization structure. This means that iteration occurs at two levels. The generalized problem to be solved in this paper is as follows:

1 Figure or Table

Cite this paper

@inproceedings{Vgh2016DeterminingAS, title={Determining Aircraft Sizing Parameters through Machine Learning}, author={J{\'a}nos V{\'e}gh and Tim MacDonald and Brian Mung{\'u}ıa}, year={2016} }