Identification of longitudinal aerodynamic characteristics of a strake-wing micro aerial vehicle by using artificial neural networks

Abstract

Appropriate modeling of unsteady aerodynamic characteristics is required for the study of aircraft dynamics and stability analysis, especially at higher angles of attack. The article presents an example of using artificial neural networks to model such characteristics. The effectiveness of this approach was demonstrated on the example of a strake-wing micro aerial vehicle. The neural model of unsteady aerodynamic characteristics was identified from the dynamic test cycles conducted in a water tunnel. The aerodynamic coefficients were modeled as a function of the flow parameters. The article presents neural models of longitudinal aerodynamic coefficients: lift and pitching moment as functions of angles of attack and reduced frequency. The modeled and trained aerodynamic coefficients show good consistency. This method manifests great potential in the construction of aerodynamic models for flight simulation purposes

Keywords

• water tunnel measurements
• neural networks
• unsteady aerodynamic characteristics
• low reynolds number aerodynamics