Windshear identification and detection in simulated and real environments
Doctor of Philosophy
This thesis considers windshear identification and detection problems by utilizing random process and statistical analysis techniques. A 3D windshear/turbulence model is developed, which covers a broad spectrum of windshear situations and can be utilized in the design of optimal and guidance trajectories. The existence of windshear patterns is justified from the analysis of simulated and real wind data. Random processes and their properties are discussed for the design of the correlation signal detectors. Numerical results demonstrate that the detection system is sensitive to the occurrence of a windshear encounter and is robust vis-a-vis noise effects. The results of this thesis make an early windshear detection possible, since computation can be performed onboard in real time. Application of the proposed detection system to the case of Flight Delta 191 leads to the conclusion that a windshear alert could have been issued 31 to 65 sec before impact, depending on the type of measurement employed in the crosscorrelation detector, specifically: 40 sec before impact for the longitudinal wind, 31 sec before impact for the vertical wind, and 65 sec before impact for the total wind. Therefore, if the Lockheed L-1011 aircraft of Flight Delta 191 had been hypothetically equipped with the present crosscorrelation detector, the pilot could have been warned 65 sec before impact, a time interval more than sufficient to avoid the crash by executing a safe abort landing maneuver.
Aerospace engineering; Statistics