e problem of building statistical models of cyber-physical systems using operational data is addressed in this paper, using the case study of aircra engines. ese models serve as a complement to physics-based models, which may not accurately re ect the operational performance of systems. e accurate modeling of fuel ow rate is an essential aspect of analyzing aircra engine performance. In this paper, operational data from Flight Data Recorders are used to model the fuel ow rate. e independent variables are restricted to those which are obtainable from trajectory data. Treating the engine as a statistical system, an algorithm based on Gaussian Process Regression (GPR) is developed to estimate the fuel ow rate during the airborne phases of ight. e algorithm propagates the uncertainty in the estimates in order to determine prediction intervals. e proposed GPR models are evaluated for their predictive performance on an independent set of ights. e resulting estimates are also compared with those given by the Base of Aircra Data (BADA) model, which is widely used in aircra performance studies. e GPR models are shown to perform statistically signi cantly be er than the BADA model. e GPR models also provide interval estimates for the fuel ow rate which re ect the variability seen in the data, presenting a promising approach for data-driven modeling of cyber-physical systems. CCS CONCEPTS •Computing methodologies →Modeling methodologies; Modeling and simulation; Model development and analysis; •Applied computing →Aerospace; Physical sciences and engineering;

Aircraft emissions are a significant source of pollution and are closely related to engine fuel burn. The onboard Flight Data Recorder (FDR) is an accurate source of information as it logs operational aircraft data in situ. The main objective of this paper is the visualization and exploration of data from the FDR. The Airbus A330 -223 is used to study the variation of normalized engine performance parameters with the altitude profile in all the phases of flight. A turbofan performance analysis model is employed to calculate the theoretical thrust and it is shown to be a good qualitative match to the FDR reported thrust. The operational thrust settings and the times in mode are found to differ significantly from the ICAO standard values in the LTO cycle. This difference can lead to errors in the calculation of aircraft emission inventories. This paper is the first step towards the accurate estimation of engine performance and emissions for different aircraft and engine types, given the trajectory of an aircraft.

Fuel burn is a key driver of aircraft performance, and contributes to airline costs and emissions. Low-altitude fuel burn and emissions, such as those that occur during climb out and approach, have a significant impact on the environment in the vicinity of airports. This paper proposes a new methodology to statistically model fuel burn in the climb out and approach phases using the trajectory of an aircraft. The model features are chosen by leveraging a physical understanding of aircraft and engine dynamics. Model development is conducted through the use of Gaussian Process Regression on a limited Flight Data Recorder archive, which also provides ground truth estimates of the fuel flow rate and total fuel burn. The result is a class of models that provide predictive distributions of the fuel burn corresponding to a given aircraft trajectory, thereby also quantifying the uncertainty in the predictions. The performance of the proposed models is compared with other frequently used Aircraft Performance Models. The statistical models are found to reduce the error in the estimated total fuel burn by more than 73% in climb out and by 59% in approach.

to initiate your request. See also AIAA Rights and Permissions www.aiaa.org/randp.

Aircraft fuel burn modeling is useful for a range of systems analysis activities. Models such as the FAA's Aviation Environmental Design Tool (AEDT) are widely used to support these activities. Previously, limited availability to thrust and movement data forced AEDT to make several simplifying assumptions, which adversely affect the accuracy of its current fuel burn estimates. This paper identifies first order airport surface fuel burn modeling enhancements in the areas of baseline taxi fuel flow modeling, taxi time estimation and pre-taxi fuel burn that may be suitable for inclusion in future versions of AEDT.