Verification of an Optimized Shape of Blended-Wing-Body Configuration using Artificial Neural Network

Abstract

In the current year’s alternative aircraft shapes, such as Blended - Wing- Body (BWB) aircraft, are considered and explored to create more effective aircraft shapes, specifically for more proficient and very huge transportation and eco-friendlier. In addition to the elimination of the tail for this specific type of aircraft and a significant reduction in equivalent weight, drag force, and radar cross-section, the available space for mounting equipment within has been improved and the operational reach has additionally been increased. Irrespective of all these stated advantages, instability is the undesirable outcome of eliminating the tail. Reviewing this deficiency involves designing a tandem of control surfaces and reflexed wing sections and utilizing a complex PC control system. Hence, several researchers have attempted to address the challenges raised by the aerodynamic shape optimization of BWBs, as well as the need to satisfy design specifications. In this paper, an experimental method was initially accepted to optimize the shape of a baseline design of a BWB. The shape was further allowed to be optimized using a Genetic Algorithm (GA). To strengthen the outcome of the optimized shape Artificial Neural Network (ANN) was used for different angles of attack ranging from -5o to 20o and airspeed ranging from 50 m/s to 700 m/s. A feed-forward back prop network with two layers of perceptron was deployed to achieve the goal of aerodynamic efficiency already set by both the experiment and CFD simulation. The goals of ANN and GA matched with a minor variation of 2% in their output results.