Development of predictive flight fuel consumption models

Abstract

Fuel consumption is a crucial consideration in the aviation industry. This last one, integral to global connectivity, faces significant environmental concerns due to the escalating demand for air travel and its substantial contribution to greenhouse gas emissions. In this thesis, we present predictive models calculating gate-to-gate fuel consumption, using simple variables such as flight distance, and taking into account the available number seats for each aircraft, in contrast with other flight consumption calculators. The main goal of this work is to construct an indicator that can be used to compare emissions with other travel alternatives. Specifically, we develop the theoretical framework for the presented models and showcase their results. Using the model with best accuracy, a LightGBM model, we demonstrate a real-world application by conducting a CO2 emission comparison between flight and rail routes.