An Efficient Split-Plot Approach for Modeling Nonlinear Aerodynamic Effects

ABSTRACT

An everyday challenge faced by experimenters across a variety of scientific disciplines is performing well-designed experiments in the presence of characteristics that pose restrictions on complete randomization of the experimental parameters. To mitigate these restrictions on complete randomization, a split-plot experimental design methodology can be employed. The current level of sophistication in split-plot designs is now sufficient to meet the demands of higher-order models in a manner straightforward enough for practitioners. A novel case of a second-order split-plot application was recently implemented in the field of aerodynamic engineering in wind tunnel testing.

Wind tunnel environments often pose restrictions on complete randomization of the test runs when aircraft physical configuration changes are required. In addition, aerodynamic empirical models require second-order effects to fit the curvature often observed in response models. Traditionally, wind tunnel testing is performed using a one-factor-at-a-time approach, which prevents capturing factor interactions and quantifying system uncertainty. This article presents a case in which a micro air vehicle (MAV) was tested in the presence of randomization restrictions with expected second-order effects utilizing an efficient design of experiments (DOE) split-plot approach.

KEYWORDS:

• design of experiments
• one-factor-at-a-time
• micro air vehicle
• response surface methods
• split-plot design

Notes

CRD = completely randomized design.

PI = prediction interval.