A bi-Level framework for aircraft design uncertainty quantification and management

Abstract

Aircraft design and development is a high-risk process. The recent obstacles with the Boeing 787 Dreamliner and Lockheed Martin F-35 show that the level of risk facing aircraft designers and manufactures has yet to be addressed. A review of work in this area reveals that methods do exist that quantify design uncertainty as well as capture common safeguards against unfavorable uncertainty realizations; however, three main capability gaps currently inhibit the effectiveness of uncertainty quantification and management: physics-based analysis, data-based uncertainty quantification, and a bi-level integrated design environment. Filling these three gaps are the contributions of this work. The geometry is explicitly modeled to retroactive changes to the design can be made in response to unfavorable uncertainty. Richardson's Extrapolation Method generates the error data needed parametric distribution fitting and correlation testing. A convergence loop based on structural weight converges two design environments of varying levels of fidelity. These three contributions are combined to form a new framework to design uncertainty quantification and management called the Reliability Assessment using Bi-level Design Analysis (RABiDA) framework.