New methods aiming to improve the performances of aircraft flight trajectory optimization algorithms

Abstract

In this thesis, three investigations aimed to explore new ways for improving the performances and capabilities of the flight trajectory optimization algorithms used by the Flight Management System. The first investigation explored a new method of selecting the geographical area considered in the flight trajectory optimization, and the construction of a corresponding routing grid. The geographical area selection method ensured the separate control over the maximal trajectory distance between the departure and destination airports, and the size of the operational area around the airports. The performances of the proposed method were analyzed using flight data from three commercial flights corresponding to short and long-haul flights. The analysis showed that the grids constructed using the proposed method had a lower number of grid nodes than the rectangular grids covering the same maximal and minimal latitudes, and longitudes. Thus, an optimization algorithm would have to evaluate a smaller number of waypoints. The analysis also showed that the proposed method was more adapted for medium and long-haul flight trajectories than for short flight trajectories. The second investigation explored a new method for reducing the volume of recurring segment performance computations, and the execution times demanded by a flight trajectory optimization algorithm. The proposed method constructed a look-up structure, defining the still-air performance parameters of the ensemble of vertical flight path segments available for the construction of the optimal trajectory. It also constructed a corresponding graph which could be used for aiding in the selection of the vertical flight path segments. The look-up structure and graph construction used the same aircraft performance model and data as the FMS trajectory computation algorithms. The following limitations were imposed in the development of this method: 1) the set of segments defined one climb and multiple horizontal constant-speed cruise, climb-in-cruise, and descent flight paths connecting the Take-Off and the End Of Descent; 2) for each flight phase, the segments correspond to a consigned speed schedule, defined as a couple of Indicated Air Speed and Mach values, and a consigned air temperature; the cruise altitudes were limited by the imposed minimal value, and by the maximal value allowed by aircraft performances, at intervals of 1,000 ft; the number of descent paths was selected through the number of aircraft Gross Weights at the End Of Descent. A number of nine test scenarios were used to analyze the performances of the proposed method, such as: 1) the number of segments composing the look-up structure; 2) the number of graph nodes; 3) the number of possible vertical flight paths connecting the Take-Off to an End Of Descent; 4) the minimal and maximal flight time and distance values,and their corresponding vertical flight paths; 5) the distribution of the vertical flight paths’ flight time versus flight distance values; and 6) the execution times required to construct the look-up structures and graphs. The third investigation explored a new method used for the geometrical construction of an optimal vertical flight plan as a function of the lateral flight plan waypoints’ along-the-track distance from the initial waypoint, their altitude and gradient restrictions, and a set of preferred gradient values defined as a function of flight phase and altitude. The main advantage of the proposed method resides in its reduced complexity, and in its increased processing speed relative to the speed of the methods employing the aircraft performance model. A second advantage is the generation of a ground-fixed optimized vertical flight plan, not affected by changing wind conditions. Two implementations of the proposed method adopting different segments construction strategies for consecutive segments leading to conflicting gradient and horizontal segment length requirements were analyzed using 48 test scenarios.

Date	19 Dec 2017
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Ruxandra Botez (Supervisor)