|
Guest >> Sign In |
|
|
DIRECTED
ENERGY
PROFESSIONAL
SOCIETY
Abstract: 24-Symp-086
|
|
|
|
UNCLASSIFIED, PUBLIC RELEASE
Nowcasting, Nearcasting, and Forecasting: Atmospheric and Optical System Performance Prediction at Various Time Scales
This presentation describes newly developed methodologies for predicting atmospheric conditions and by extension the anticipated optical system performance. These predictions can aid in DE test planning, determining the anticipated effectiveness of DE, and automating beam control settings. It is well known that optical turbulence can vary several orders of magnitude over the course of one day. The time windows surrounding the morning and evening neural events can experience those large variation of turbulence in the course of just a few hours. At those hourly time scales, NWP modeling is able forecast the weather conditions that drive the optical turbulence. High temporal resolution for NWP modeling is typically hourly, and thus to prediction conditions over the next few minutes or few seconds new techniques most be explored. Nearcasting is a new turbulence prediction methodology that does not rely on NWP. Instead, nearcasting observes the recent trend of weather and measured turbulence to predict the incoming trend through extrapolation. A machine learning approach is used to train a model that learns the relationship of the prior several hours to predict the next few minutes. In a similar way, nowcasting relies on real-time high speed continuous measurements to predict the future few milliseconds. We demonstrate nowcasting by making high speed wavefront measurements and predicting multiple future steps of the incoming wavefront.
UNCLASSIFIED, PUBLIC RELEASE
|