Data-Driven Method for Flow Sensing of Aerodynamic Parameters Using Distributed Pressure Measurements

The real-time identification of inflow aerodynamic parameters such as the flow separation situation, angle of attack, and inflow velocity is challenging. In this paper, a new data-driven strategy is proposed to attain real-time identification of the inflow aerodynamic parameters through a combination of experimental data (offline) and distributed pressure sensor measurements (online). In the offline procedure, pressures on the airfoil surface are measured by 10 distributed sensors under 45 different conditions. Particle image velocimetry measurements are recorded to determine the correlation between the pressure distribution and the aerodynamic parameters. Proper orthogonal decomposition (POD) is applied on the pressure data under all conditions to encode this correlation into a three-dimensional domain, with a compression ratio of more than 90%. In the online procedure, the k-nearest neighbor algorithm is used to identify the aerodynamic parameters of testing data in the established POD domain. Furthermore, a mathematical model is applied to optimize the pressure sensor locations. Using just four pressure measurement points, accuracy of 100% can be achieved for the flow separation detection, angle of attack, and inflow velocity. After flow separation, the new approach achieves an error of approximately +/- 1 deg for the angle of attack.

Automated summary

This paper introduces a new data-driven method combining experimental data and distributed pressure sensor measurements to achieve real-time identification of inflow aerodynamic parameters. Through processing and optimizing large amounts of data, the accuracy and efficiency of parameter identification are improved significantly.