Aircraft fuel models allow operators to predict the amount of fuel an aircraft will consume at a specified flight condition. These models can then be used to generate fuel tables or as standalone predictors for flight planning software to estimate the amount of fuel an aircraft will use during a flight. This Paper proposes four fuel models that are generated based on recent historical high-frequency flight data to predict the fuel use of individual aircraft. Two parametric models are described: a novel dynamics-based parametric model characterized by estimating the aerodynamic and propulsion properties of the aircraft and a simple linear model. Two nonparametric models, a Gaussian process regression model and a -nearest neighbors regression model, are also described. The models are tested using a set of flight data consisting of flight records from 40 months of time-partitioned training and test sets for each of three individual airframes, each airframe being a different aircraft type, including both narrow- and wide-body jet transport aircraft. The dynamics-based parametric model is demonstrated to consistently improve prediction accuracy for climbing and cruising flight over the existing fuel prediction method.
Using High-Frequency Data for Predicting Fuel Use of Jet Transport Aircraft
Journal of Aircraft ; 54 , 6 ; 2115-2125
2017-09-11
11 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Using High-Frequency Data for Predicting Fuel Use of Jet Transport Aircraft
Online Contents | 2017
|Long range transport aircraft using hydrogen fuel
Tema Archiv | 2013
|Fuel for future transport aircraft
Tema Archiv | 1983
|Fuel reserves for jet transport aircraft
Engineering Index Backfile | 1954
|