A mission fuel performance model based on hybrid flight physics and QAR data

Abstract

Use of operational data such as those from QAR (Quick Access Recorder) has recently attracted interest in building high-accuracy flight fuel models. This is often combined with applying some machine learning algorithms to improve the model’s fidelity. However, the data-based approach lacks the physical characteristics of the aircraft flight performance models and is challenging to interpret and use in optimizing aircraft designs. This paper proposes a collaborative optimization process based on a physics-based aircraft multidisciplinary sizing tool and a data model built from flight data. First, an enhanced aircraft sizing tool is used to provide initial estimation of the aircraft design parameters based on the top-level requirements. Unknown parameters in the sizing model are determined using data-based approach which include both aircraft operational and flight parameters. Aircraft operational parameters include actual passenger weight, cargo weight, fuel weight, cruising Mach number, and other essential operational parameters. Aircraft flight parameters include information on aircraft, route, and weather etc., derived from QAR data and open-source flight databases. Aircraft design, operation, and flight parameters are coupled with an aircraft performance model, which can be used in a collaborative multi-parameter optimization framework to optimize aircraft design and operations for improved fuel performance.