The Use of an Integrated Optimization Framework for Optimizing Noise and Fuel

Abstract

Travelling by air has become commonplace, and more and more people can afford themselves to fly. In twelve years time, the amount of people that were carried by airlines doubled to four billion passengers in 2017. At Amsterdam Airport Schiphol, the main gateway into the Netherlands and also a large hub for connecting flights, saw a growth of 20 million passengers in four years time. The rise in air travel also has its downsides. Not only the boundaries of airport and airspace capacity are run into, noise annoyance and pollution caused by aircraft are also everyday's business. The growth in aviation and population will inevitably lead to conflicts. Several strategies have been proposed in literature and by legislators to address this conflict. Advanced aerodynamics, aircraft engines and alternative fuels can be considered as long term solutions for this, but changing operational procedures of aircraft and airport is considered as a good solution in terms of costs and time-wise. Most research on changing operational procedures has shown significant results, however only consider a single operation while these are of influence on one another. The integration of several operational procedures into a single problem has been recognized as one of the future research topics. In this research an integrated optimization framework is proposed that can simultaneously optimize aircraft and airport operations by allocating routes, flight procedures and runways. The framework is able to handle fuel usage, noise annoyance and capacity. Amsterdam Airport Schiphol will be used as case study airport, and a real flight departure schedule is used as reference. The framework consists of two main steps. In the first step, flights are groups according to their aircraft type, destination and departure time, and are subsequently distributed over the available runways, routes and departure procedures. This distribution is performed with a \acrshort{NSGA}-II algorithm that is capable of handling non-linear, multi-objective optimization problems. In the second step, the flight distribution is apportioned over the flights in the schedule, while adhering the separation requirements. In this way it is verified if the required capacity is available, by adhering the separation at departure, a possible crossing point and the terminal point between consecutive flights. The second step is implemented with a \acrshort{ILP} algorithm. Results from the case study has shown that significant gains can be obtained in terms of noise annoyance. While requiring little extra fuel, up to 4\%, the noise annoyance can be lowered with 31\%. Not much variance is possible for flights that operate during the night hours, and so all gains are obtained on flights in the daytime. Especially the amount of people that are highly annoyed can be lowered of up to almost 50\%. A further analysis of the noise contours shows that the framework is able to shift noise in such a way that the noise levels drop below the threshold values of the noise equivalence criteria. In other words, the noise contours can decrease in size and with that the amount of people annoyed is lowered. These results are obtained by shifting flights from runway 18L to runway 24, which is more noise friendly, and by using \acrshort{NADP}2 more frequently compared to \acrshort{NADP}1. The model is also able to make a distinction between destinations and aircraft type, and puts heavy aircraft more frequently on shorter flight legs in order to minimize the fuel usage and the exposure to noise of residents. An examination of the entire solution set in terms of the separation requirements indicates that most of the found results are feasible. By shifting more flights to a single runway, the required capacity is at stake and adding extra flights with delay will cause an exponential growth in total required delay.
Using the integrated optimization framework leverages the potential of aircraft specific capabilities by allocating the right runway, route and procedure to each flight in such a way that an optimum noise contour is formed. Most gains can be obtained for the number of people annoyed and disturbed, and as these noise equivalence criteria are directly related to the well-being of residents around the airport and their opinion on the operation of the airport, this can only be considered as beneficial. The available capacity of a runway is deployed to alleviate noise annoyance by residents. The framework proposed in this research can be improved at several components in future research. The addition of arrivals, extra routes and procedures would extend the applicability and value of the framework. Another improvement that can be made is the incorporation of the second step into the first step. This would eliminate infeasible solutions inside the solution set.