A PCA-ANN-based inverse design model of stall lift robustness for high-lift device

The concept of stall lift robustness for high-lift device (HLD) measures the stability of lift values under a series of angles of attack (AoA) around stall, which plays a significant role in flight safety. In this article, a stall lift robustness design with the consideration of aerodynamic constraints (stall AoA, average lift, etc.) is carried out, where design targets are set as a series of lift values on Lift-AoA curve. A Principle Component Analysis (PCA)-Artificial Neutral Network (ANN)-based inverse design model is introduced. The design targets are transformed by PCA for data dimension reduction. Then, the new set of design targets are input into the surrogate model of ANN, and corresponding geometry of new HLD is predicted. The ANN is constructed through database and sample points are screened considering lift unsteadiness. The design procedure is iterated to meet the design accuracy. The process of stall lift robustness design with the proposed model is discussed in this article, and the design results are validated by Detached-Eddy Simulation (DES).