Parametric estimation scheme for aircraft fuel consumption using machine learning

The most efficient technique that is used for aircraft engine tuning is through mounting the engine on the engine test bench (ETB) to analyze, tune and monitor its variables through the ETB run. It is practically very difficult to unmount the engine from the aircraft and mount it on the ETB for analyzing and estimating a single variable such as fuel consumption or oil temperature as the unmounting process requires huge manpower and machinery. This problem can be resolved if the fuel consumption of an air vehicle is estimated without unmounting the engine from the aircraft through applying data analytics and machine learning models. Therefore, in this paper, the fuel consumption of an aircraft is analyzed and estimated through advanced data science techniques. The dataset went through data analyzing and preprocessing techniques before applying multiple machine learning models such as multiple linear regression (MLR), support vector regression, decision tree regression and deep learning algorithm RNN/LSTM. The performance of algorithms has been evaluated using model evaluation methods such as mean absolute error (MAE), root mean square error (RMSE) and coefficient of determination. The models are evaluated in taxi, cruise and approach flight phases where the LSTM performs excellent among all other algorithms with RMSE 15.1%, 10.5% and 0.9%, respectively.