Wind-shear prediction with airport LIDAR data

Dangerous weather is an important factor of flight safety. Particularly, wind-shear is the most dangerous weather. In this paper, four traditional methods (Grey model, BP neural network, Brown three exponential smoothing, and Support vector regression) on PPI scan data are used in wind field forecast experiments, from which forecast wind speed map can be got. We first use the above four methods to forecast wind field with glide path scan data and extract headwind and wind-shear ramp from the data and show the wind-shear alert. Then, a new method named grey forecast with Position Amendment and Fluctuation Compensation (PAFC) is proposed, which employs BP neural network as the position amendment module and Brown three exponential smoothing as the fluctuation compensation module. The experiment results on HKIA Doppler LIDAR data show the good performance of our method.