Skip to main content.
Advanced search >
<< Back to previous page Print

<< Friday, March 24, 2017 >>

Remind me

Tell a friend

Add to my Google calendar (bCal)

Download to my calendar

Bookmark and ShareShare

Changing Fuel Loading Behavior to Improve Airline Fuel Efficiency

Lecture | March 24 | 4-5 p.m. | 290 Hearst Memorial Mining Building

Lei Kang, UC Berkeley

Institute of Transportation Studies

Abstract: Airlines rely on flight dispatchers to perform the duty of fuel planning. The required trip fuel is calculated by airlines’ Flight Planning Systems (FPS). However, the FPS trip fuel predictions are not always accurate. If planned trip fuel is higher than actual trip fuel, then a flight will waste fuel by carrying excess fuel weight. On the other hand, if trip fuel is under-estimated, then a flight might run into fuel emergency. In practice, dispatchers may also load contingency fuel to mitigate the risks of under-prediction. FPS also calculates recommended contingency fuel quantity for dispatchers called statistical contingency fuel (SCF). However, dispatchers will almost always load extra fuel above suggested SCF values. Therefore, airline fuel efficiency can be improved by more accurate fuel predictions, a deeper understanding of dispatchers’ fuel loading behavior, and more reliable SCF recommendations. Based on a large scale flight fuel loading dataset provided by a US major airline, an ensemble learning algorithm is proposed to improve fuel burn prediction. This method is found to reduce prediction error by over 50% compared to airline’s own predictions. By merging with a dispatcher survey, we are able to integrate dispatchers’ latent attributes into contingency fuel loading modeling. Furthermore, random quantile forests method will also be discussed in improving SCF recommendations. The benefit of improved fuel efficiency will be measured by estimating cost-to-carry reduced unnecessary fuel loading.

Bio: Lei Kang is a Ph.D. candidate of the Institute of Transportation Studies in the Department of Civil and Environmental Engineering, University of California, Berkeley. He received a Master of Arts degree in Biostatistics from the Division of Biostatistics at University of California, Berkeley. He also obtained his Master’s degree in Transportation and Infrastructure Systems Engineering from Purdue University. Lei's Bachelor’s degree is in Transportation Engineering from Tongji University in Shanghai, China. He is a member of the Committee on Airfield and Airspace Capacity and Delay, Transportation Research Board. His research interests are in the application of statistical methods and machine learning techniques to air traffic management and airline fuel loading decisions. He is also interested in causal inference in the area of traffic safety.