Aircraft Characterization in Icing Using Flight Test Data

The University of Illinois, in conjunction with NASA Glenn, has conducted a flight test program to investigate aircraft icing effects and develop aircraft icing effects characterization techniques. Flights were conducted in 2001 and 2002 to collect data in clear air as well as in natural icing conditions. These data were used to identify the effects of icing on aircraft performance and control. It was also used to aid in the development of an H ∞ characterization algorithm being designed specifically for the Smart Icing System (SIS). Flight test data analysis was accomplished using a modified stepwise regression technique. Systems IDentification Programs for Aircraft (SIDPAC) was used for the linear regression analysis. The results of this analysis revealed multiple parameters that clearly indicated both the presence and severity of the ice accretion. Atmospheric turbulence was shown to significantly affect this method of parameter identification. A real-time H ∞ parameter identification (PID) algorithm has previously been developed as part of the icing characterization function of the Smart Icing System. This paper presents an initial validation of the H ∞ PID algorithm wherein the algorithm is applied to the flight test data after the fact, but in a manner that is consistent with the ultimate real-time application. This mock real-time validation demonstrates that the H ∞ PID algorithm provides pitching moment derivative estimates that are consistent with those provided by SIDPAC for both clear-air and icing conditions in the absence of turbulence. As with the stepwise regression technique, the H