SIMPLIFIED MATHEMATICAL MODEL OF SMALL SIZED UNMANNED AIRCRAFT VEHICLE LAYOUT

• Published in: Nauchnyĭ vestnik MGTU GA (Online) Vol. 19; no. 6; pp. 86 - 94
• Main Authors: V. V. Vyshinsky, A. O. Kislovskiy, S. A. Kolchev
• Format: Journal Article
• Language: Russian
• Published: Moscow State Technical University of Civil Aviation 01.01.2017
• Subjects: aircraft layout; artificial intelligence; artificial neural network; computational fluid dynamics; mathematical model; small-sized unmanned aircraft vehicle
• ISSN: 2079-0619; 2542-0119

Strong reduction of new aircraft design period using new technology based on artificial intelligence is the key problem mentioned in forecasts of leading aerospace industry research centers. This article covers the approach to development of quick aerodynamic design methods based on artificial intelligence neural system. The problem is being solved for the classical scheme of small sized unmanned aircraft vehicle (UAV). The principal parts of the method are the mathematical model of layout, layout generator of this type of aircraft is built on aircraft neural networks, automatic selection module for cleaning variety of layouts generated in automatic mode, robust direct computational fluid dynamics method, aerodynamic characteristics approximators on artificial neural networks.Methods based on artificial neural networks have intermediate position between computational fluid dynamics methods or experiments and simplified engineering approaches. The use of ANN for estimating aerodynamic characteris-tics put limitations on input data. For this task the layout must be presented as a vector with dimension not exceeding sev-eral hundred. Vector components must include all main parameters conventionally used for layouts description and completely replicate the most important aerodynamics and structural properties.The first stage of the work is presented in the paper. Simplified mathematical model of small sized UAV was developed. To estimate the range of geometrical parameters of layouts the review of existing vehicle was done. The result of the work is the algorithm and computer software for generating the layouts based on ANN technolo-gy. 10000 samples were generated and the dataset containig geometrical and aerodynamic characteristics of layoutwas created.