Impact of Designed Experiments on Efficiency, Accuracy, and Hypothetical Test Costs Using a Light Aircraft Wing Model
Stephen Szaruga, John H. Doty · 49th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition · 2011
Preliminary light aircraft design is often conducted using simulated models to provide guidance for potential candidates suitable for more advanced analysis. Herein we study a 10-variable light aircraft wing model using two different approaches: a parametric search and statistically-driven Design of Experiments (DoX). In an optimization study, the DoX approach yielded the same minimum wing weight, to within 1% accuracy, but required as few as 98% fewer test points than the exhaustive search. For the simulation study conducted, the DoX approach required only 5 of the 10 variables in the original model to achieve these results, indicating that 5 of the variables are not required. A hypothetical cost model is postulated by assuming that each of the design points requires a quantifiable time and cost to implement. In this manner, the impact of the more efficient DoX approach is cast in economic terms with a much more powerful message.