Large-scale MDO under uncertainty of an eVTOL aircraft using dimension reduction via global sensitivity analysis

Electric vertical takeoff and landing (eVTOL) aircraft have gained significant attention from both industry and academia due to their potential to transform urban transportation and promote sustainability through zero-emission propulsion. Recent work has demonstrated the effectiveness of multidisciplinary design optimization (MDO) in enhancing eVTOL performance at the conceptual design stage. However, most existing studies have focused on deterministic MDO problems neglecting the impact of real-world uncertainties. To address this gap, we formulate and demonstrate the solution of the first large-scale MDO under uncertainty (MDOUU) problem for NASA’s Lift-Plus-Cruise configuration. This problem encompasses 51 design variables and 7 independent uncertain inputs. To keep the computational cost tractable, we introduce a novel sensitivity analysis-based, shared-grid uncertainty quantification framework that (i) screens each quantity of interest (QoI) with first-order Sobol indices and (ii) evaluates the resulting low-dimensional quadrature rules on a single, unioned node set. Numerical results demonstrate that the proposed UQ strategy achieves accurate statistical estimates for all 7 MDOUU quantities of interest (QoIs), with errors all below 5% using only 9 model evaluations. This strategy enables the efficient solution of the MDOUU problem, resulting in a reduction of 10.4% in aircraft empty weight compared to a baseline MDO design that is optimized using conservative safety factors.