Delivering Program Efficiency to Aerospace Testing Using Designed Experiments Open Access
Downloadable ContentDownload PDF
Given the increasing complexity of systems and the cost associated with test and evaluation of aerospace systems, more efficient methods are sought. Randomized test designs for aviation developmental test activities and other complex systems may not enable safe test conduct and may be prohibitively costly from a financial or time point of view. This research reviews Design of Experiments (DoE) test design approaches applicable to aerospace prototype test and evaluation activities. It proposes the use of Split Plot Optimal Designs to leverage advantages of DoE while satisfying requirements for limited randomization of the test runs. Through the use of case studies, the Split Plot Optimal Design approach is demonstrated to provide a 58% cost and schedule savings versus a One Factor At a Time approach, and 53% savings from the fully randomized Central Composite Design, while maintaining relevant statistical power. Through the use of Monte Carlo data simulation, the designs are evaluated for application to linear and quadratic models, with statistically significant results measured by Chi Squared and Kolmogorov-Smirnov tests.