Comparison of Aero Data Model Quality and Cost for Design of Experiments and One-Factor-at-a-Time Test Matrix Design Approaches

摘要

A posttest evaluation was performed to investigate the possibility of using a Design of Experiments (DOE) approach for test matrix development and posttest response surface models (RSM). For this study the Arnold Engineering Development Center (AEDC) provided a historical data set from a wind tunnel test of a scaled missile tested at transonic conditions. This historical data set containing roughly 6,600 data points was used to create "Truth" RSMs for all six of the aerodynamic force and moment coefficients. An augmented nested face-centered design (NFCD) was used to sample the Truth models, and third-order polynomial RSMs were created from the NFCD design points. Comparisons between the Truth and NFCD RSMs, and the estimated cost and schedule of each design, were examined.