Correlation Analysis between Airfoil Icing and Icing Cloud Conditions Based on PLS Methods

Ice shape feature parameters include iced area, icing mass, lower horn angle, upper horn angle, lower icing limit, upper icing limit, lower horn thickness and upper horn thickness, etc. Using the Partial Least Square (PLS) method, regression models for the first four parameters are established, where the first two are presented in this paper and the other two are not due to the limitation of space; while models for the other four ice features parameters are not established successfully and some reasons are given. Variation trends of the first two parameters predicted by PLS regression models built are compared with those obtained byMiller. It shows the two agree quite well. Air Temperature (T) and Liquid Water Content (LWC) has the most effects on iced area. Median Volume Diameter (MVD), T and Freezing Fraction (h ) has the most effects on icing mass. Other icing cloud condition parameters have little effects on them.