New Methodology for Robust-Optimal Design with Acceptable Risks: An Application to Blended/Wing/Body Aircraft

In this paper, a new design methodology guaranteeing optimal performance and constraint satisfaction with acceptable risks is presented. This methodology is particularly tailored to conceptual design, wherein variability stemming from modifications that will inevitably occur in downstream design phases must be taken into account. Based on second-order Taylor expansions, a new method allowing the transformation of this robust-optimal design problem with acceptable risks into an equivalent deterministic quadratically constrained quadratic programming (QCQP) problem is provided and then applied to an example problem for validation. Hence, all the available QCQP optimizers can be used to find a robust-optimal solution. The proposed methodology is applied to the blended-wing/body (BWB) at the conceptual design phase, where robust drag minimization is sought under a pitching moment constraint. Numerical results show the impacts on both the BWB optimal planform parameters and guaranteed drag as a function of the acceptable risk levels taken. The study shows that two of the six planform shape variables are significantly affected by the risk level, namely, the kink parameter and the outer wing span. Variations of almost 20% are obtained on these two variables for a risk level of 0.1% as compared to a standard deterministic optimization.

Automated summary

This paper presents a new design methodology that guarantees optimal performance and constraint satisfaction with acceptable risks. It is specifically tailored to conceptual design, where variability arising from modifications in downstream design phases must be considered. The methodology utilizes second-order Taylor expansions to transform the robust-optimal design problem into an equivalent deterministic quadratically constrained quadratic programming (QCQP) problem. The proposed methodology is applied to a blended-wing/body (BWB) design problem for validation, showing the impacts of acceptable risk levels on the optimal design parameters and guaranteed drag.