Aerodynamic Design Optimization of a Staggered Rotors Octocopter Based on Surrogate Model

Improving aerodynamic performance with limited dimensions is becoming more important in the design of multirotor aircraft. Complex coupling interactions exist between the structural parameters and aerodynamic performance of the double-layer staggered rotor octocopter. This paper develops an integrated aerostructural design optimization approach based on a surrogate model to reduce the high computation cost of computational fluid dynamics (CFD) simulation experiments, where the surrogate model is constructed by using the Latin Hypercube design, and different approximation methods are compared. Aiming to improve the global accuracy of the surrogate model, computer-aided design (CAD) modeling and CFD simulation experiments were carried out according to the sample points designed by the Latin Hypercube method. A radial basis function surrogate model was constructed based on 63 sample points obtained with CFD simulations after comparison with other approximation methods. The optimization for aerodynamic/structural design was formulated and solved by the multiisland genetic algorithm. The results show that the surrogate model has an accurate predictive ability, and the optimal solution obtained from the optimization has a better aerodynamic performance than the samplings. Compared with the initial optimum data, the optimum solution proposed in our study could generate 102.7% more thrust, and the objective function $y$ is improved by 105.1% according to the CFD simulation results. The approximation and optimization approach effectively reduces the cost of many CFD calculations in aircraft design and provides a global prediction for the performance of the octocopter.