Online identification of aerodynamics with fast time‐varying features using Kalman filter

Abstract

Abstract This paper investigates the online identification method of the unsteady aerodynamics in the post‐stall manoeuvres. The main unsteady features are hysteresis and fast time varying due to the complex flow structure at large angles of attack. When limited test data are available, the data‐driven model cannot be identified accurately, especially in dealing with the fast time‐varying features, so online identification is required. A constant acceleration model‐based Kalman filter is constructed by taking into account the unsteady item in the aerodynamic model. Unlike the previous work which only uses the information of second‐order moments of errors, the proposed method uses the maximum correntropy criterion. Simulation studies according to the wind tunnel test data show that the proposed method can effectively track the fast time‐varying features and estimate the aerodynamic coefficients with high accuracy.

Keywords

• Applied fluid mechanics
• Filtering methods in signal processing
• Fluid mechanics and aerodynamics (mechanical engineering)
• Other topics in statistics
• Other topics in statistics