Impact of the Number of Wing Stations on the Twist Distribution of a BWB Conceptual Aircraft Design in a MDAO Environment to Match Target Lift Distribution

Abstract

The design of a blended wing body (BWB) has been the subject of numerous research projects worldwide, as it offers significant advantages over a conventional tube-and-wing configuration. The European Clean Sky 2 project NACOR investigated the potential of innovative unconventional aircraft architectures. A BWB configuration was identified to reducing fuel consumption by 10% compared to a relevant reference aircraft with entry into service 2035 and prevailed against box-wing or strut-braced wing aircraft. For this reason, the need to be capable of designing and evaluating BWB configurations was recognized. Especially for the design of BWBs, an optimized wing twist is important to achieve high aerodynamic performance, but on conceptual design level the number of wing stations is often limited to reduce complexity which impedes this optimization. Computationally expensive calculations such as Reynolds-averaged Navier-Stokes (RANS) or Euler method are required to optimize wing twist, but this reduces the number of possible designs examined in a given time. Therefore, this paper presents the feasibility of flexible number of wing stations to achieve an aerodynamically optimized wing twist for BWB architectures on conceptual aircraft design level with embedded low fidelity multi-lifting line approach. The aim of the flexible wing modelling combined with a multi-disciplinary design analysis and optimization (MDAO) aircraft design environment is to exploit the benefits of low computational costs and high flexibility with respect to the vehicle architecture to be analyzed with an adequate level of fidelity. The resulting fast analysis capability enables a wide range of possible sensitivity studies and optimizations to ensure a more robust aircraft design space exploration. The target lift distribution is determined by high fidelity RANS calculations and used as an objective function for the twist optimization. In addition, typical lift distributions, i.e. elliptical and triangular lift distributions, are used to benchmark the results. The final results show the influence of the number of wing stations on the resulting twist distribution, induced drag as well as the CPU cost, and provide a proposal for an appropriate number of wing stations to be used.