Data Intelligence & Fuel Efficiency: A Data-Driven Approach to Manage Uncertainties in Flight Fuel Planning for Airlines

Project: Research

View graph of relations


The airline industry has been striving to improve fuel efficiency. The incentives are not only economic - fuel cost accounts for a third of the operational cost for most airlines – but also environmental, since the industry is under growing pressure to reduce carbon emission and to promote sustainable growth. Existing approaches on fuel efficiency focused on aircraft or engine designs, flight operations, e.g. optimal altitude and speed, and maintenance procedures.Flight fuel planning emerged as a new area for additional fuel savings. Each aircraft carries extra fuel in order to prevent deplete given the uncertainties of fuel consumptions. Beyond the reserve fuel mandated by regulations, airlines load Contingency Fuel at their discretion. The amount of Contingency Fuel is determined by the single highest fuel consumption record in the past. As a result, most flights are carrying much more fuel than needed, and the cost of carrying these “dead weight” is significant. If fuel consumption can be more accurately predicted, the need for Contingency Fuel can be reduced for the benefit of fuel savings across all flights. With the rapid development of data intelligence and the large amount of operational data accumulated at airlines, a data-driven approach will change how uncertainties are managed for fuel planning.In this research, we propose a data-driven approach to better manage uncertainties in flight fuel consumption for airlines. The aim is to build a predictive model that can better estimate each flight’s fuel consumption by combining multiple data sources (e.g. traffic conditions, Flight Data Recorder data, etc.) which are not typically included in the current practice. Three research tasks are scheduled to achieve the aim: 1) analyze discrepancies between planned and actual fuel consumptions and identifying key contributing factors; 2) develop a data-driven model to predict flight fuel consumption; 3) optimize flight fuel loading towards a data-driven fuel planning strategy.The project is in line with the PI and Co-I’s research experiences in air transportation systems, data analytics, and optimization. A pilot study was conducted in collaboration with a major airline company based off Hong Kong. We take encouragement from preliminary results which suggest promising agreement between actual fuel consumption and prediction using our data-driven approach. A successful completion of this project will contribute to academic literature in data intelligence and aviation fuel efficiency. The collaborative nature of this project between academia and industry may impact flight fuel planning practice and further promote fuel efficiency. 


Project number9042910
Grant typeGRF
Effective start/end date1/01/20 → …