Design and optimization of a blended wing body UAV using AI-driven surrogate modeling and CFD analysis

Abstract

This paper presents a practical, end-to-end approach to improving the flight endurance of a one-meter wingspan blended-wing-body (BWB) unmanned aerial vehicle (UAV) powered by a low-power propulsion system. It aims to increase aerodynamic efficiency and confirm structural viability for small-motor, energy-limited missions. The study combined high-fidelity computational fluid dynamics (CFD) to evaluate aerodynamic behavior with machine learning surrogate modelling and Bayesian optimization to search the design space efficiently. Key geometric variables i.e. wing planform, airfoil choice, winglet geometry and surface twist were evaluated using a Gaussian Process Regression surrogate trained on CFD results, enabling rapid prediction of aerodynamic performance across many candidate configurations. Optimized surface geometries increased the lift-to-drag ratio significantly, reducing required propulsion power and extending predicted endurance. To ensure the design could withstand operational loads, the airframe was sized for manufacture using a twill weave 3K 200gsm carbon fiber/epoxy composite and evaluated with finite element analysis (FEA); results showed the reinforced structure met safety and performance criteria for expected flight loads. A prototype was fabricated and subjected to wind tunnel testing, which demonstrated longer-duration flights under lower wind speed, as to propulsion, compared with baseline configurations. Together, the aerodynamic and structural optimizations produced a lightweight, manufacturable BWB UAV suitable for energy-efficient tasks such as long-range logistics, surveillance, and emergency response where low propulsion power is a constraint. The methods and results offer a repeatable workflow for integrating CFD, surrogate modelling, Bayesian design search, composite design, and physical testing to develop low-power UAVs, and provide practical guidance for designers and engineers working on similar low-energy aircraft systems.

Keywords

blended wing body UAV aerodynamic optimization surrogate modeling Gaussian process computational fluid dynamics (CFD)composite fabrication

Get full access to this article

View all access options for this article.

References

Liebeck

. Design of the blended wing body subsonic transport. J Aircraft 2004; 41(1): 10–25. https://doi.org/10.2514/1.9084

Wang

Zhou

. Aerodynamic design, analysis and validation of a small blended-wing-body unmanned aerial vehicle. Aerospace 2022; 9(1): 36. https://doi.org/10.3390/aerospace9010036

Clainche

Ferrer

Gibson

, et al. Improving aircraft performance using machine learning: a review. Aero Sci Technol 2023; 138: 108354. https://doi.org/10.1016/j.ast.2023.108354

Boutemedjet

Samardžić

Rebhi

, et al. UAV aerodynamic design involving genetic algorithm and artificial neural network for wing preliminary computation. Aero Sci Technol 2018; 84: 464–483. https://doi.org/10.1016/j.ast.2018.09.043

Panagiotou

Dimopoulos

Dimitriou

, et al. Quasi-3D aerodynamic analysis method for blended-wing-body UAV configurations. Aerospace 2021; 8(1): 13. https://doi.org/10.3390/aerospace8010013

Brunton

Kutz

Manohar

, et al. Data-Driven aerospace engineering: reframing the industry with machine learning. AIAA J 2021; 59(8): 2820–2847. https://doi.org/10.2514/1.j060131

Zadeh

Sayadi

. An efficient aerodynamic shape optimization of blended wing body UAV using multi-fidelity models. Chin J Aeronaut 2018; 31(6): 1165–1180. https://doi.org/10.1016/j.cja.2018.04.004

Simpson

Poplinski

Koch

, et al. Metamodels for computer-based engineering design: survey and recommendations. EWC 2001; 17: 129–150. https://doi.org/10.1007/PL00007198

Sobol

. Global sensitivity indices for nonlinear mathematical models and their monte carlo estimates. Math Comput Simulat 2001; 55(1–3): 271–280. https://doi.org/10.1016/s0378-4754(00)00270-6

10.

Menter

. Two-equation eddy-viscosity turbulence models for engineering applications. AIAA J 1994; 32(8): 1598–1605. https://doi.org/10.2514/3.12149

11.

Forrester

AIJ

Sóbester

Keane

. Engineering design via surrogate modelling, 2008. https://doi.org/10.1002/9780470770801

12.

Rasmussen

Williams

CKI

. Gaussian processes for machine learning. The MIT Press eBooks, 2005. https://doi.org/10.7551/mitpress/3206.001.0001

13.

Jones

Schonlau

Welch

. Efficient global optimization of expensive black-box functions. J Global Optim 1998; 13: 455–492. https://doi.org/10.1023/A:1008306431147

14.

Putra

Setiawan

Sehono

, et al. Analisis pengaruh metode vacuum bag terhadap geometri sayap UAV skywalker. SJME Kinematika. 2023; 8(1): 36–43. https://doi.org/10.20527/sjmekinematika.v8i1.244

15.

Pinho

Iannucci

Robinson

. Formulation and implementation of decohesion elements in an explicit finite element code. Composites Part a Applied Science and Manufacturing 2005; 37(5): 778–789. https://doi.org/10.1016/j.compositesa.2005.06.007