Formation Flight for Commercial Aviation

A Case Study


Abstract

The need for commercial aviation to be more efficient and reduce its impact on the environment is an ongoing challenge. Formation flight has the potential to significantly reduce aircraft fuel consumption by allowing `follower' aircraft to fly in the aerodynamic wake of a `leader' aircraft. However, this requirement for flights to be in close proximity for large parts of their flights, raises questions about it suitability to the range of existing flights and varied geographical regions.
Therefore in this work we explore the potential for implementing formation flight for commercial aviation for distinct case studies: Long-Haul Airline (LHA), TransAtlantic Airline (TAA), and Low-Cost Airline (LCA). Each case study represents different airline characteristics and regions typical of today's flights. Fuel-saving results for average percentage savings against solo flight are shown to be very promising, with TAA achieving almost 9%, LHA just over 6% and LCA just under 2%. This potential for fuel saving could amount to hundreds of millions of dollars in fuel savings and millions of tonnes of reduction in resulting Co2 emissions.

Methodology

Using Matlab to first find the optimal formations of size two we look at the results to see how the multi-variate results interact. Using the power of dc.js, crossfilter.js and Google Maps' Javascript Api we can quickly render the results.

Questions

We can interrogate the data to try and answer some simple questions.
• How much Co2 can we save by flying in formation?
• How does time-of-day correlate with saving?
• Can Formation Flight still be benifical for short haul flights?
• Do Formations with less deviation perform better?
• What impact does restricting changes in take-off time effect savings?

Papers

Optimal routing and assignment for commercial formation flight
T. Kent
2015
Thesis
Department of Aerospace Engineering
University of Bristol
Analytic Approach to Optimal Routing for Commercial Formation Flight
T. Kent & A. Richards
2015
AIAA Journal of Guidance, Control and Dynamics
Volume 38, No. 10
On optimal routing for commercial formation flight
T. Kent & A. Richards
2013
AIAA Guidance, Navigation, and Control Conference
Boston, MA

Choose a dataset

Case Study: Transatlantic Flights

This dataset is of the standard 210 transatlantic flights flying from the US to Europe. These are currently without a particular date and so are assumed to be all on the same day. All 22k of the combinations of the 210 flights are evaluated. The datasets available are for those which have then been assigned into formation via a MILP - under the noted constraints, and the dataset of the unassigned formations.
MILP assigned formations - with constraints
Using only the combinations whose take off times only need to be altered by a certain threshold

Unassigned formations
All combinations of flights evaluated for a given day

Case Study: Low Cost Airline

This dataset contains 4238 different flights routes from a European low cost airline over 7 days, some flights running on multiple days meant that the total individual flights totalled 8750. As flights repeat over the week and multiple identical routes are flown on the same day (but at different times) the results need to be filtered.

Firstly the data is broken down into individual days (roughly 1000 flights each day), then all possible combinations (roughly 750k each time) are evaluated for formation. The unassigned formations dataset are then all the favourable formations from all combinations. These give a broader overview of all combinations and the spread of results. The MILP assigned formations, are then those which maximise the total possible saving. These are run for each day, and for different thresholds of alterations to takeoff time. The change in Take off time is important, as multiple identical flights on a day would quickly be put together otherwise, producing unrealistic results.
MILP assigned formations - with constraints
Using only the combinations whose take off times only need to be altered by a certain threshold

Unassigned formations
All combinations of flights evaluated for a given day

Case Study: Long-Haul Airline

This case study represents a major airline company based in Singapore, serving flights from Southeast, East and South Asia to many domestic and international destinations. It acts as a good case study for a wide range of flight distances, ranging mostly between medium-haul to super-long-haul flights. The main difference between these flights and ones seen in the other case studies is that the vast majority of the routes are either flying from or to Singapore acting like a hub-network so not all routes are natural candidates for formation flight. The dataset contains 417 different flights running over a 7 day period. The number of flights per day ranges between 228 and 242 resulting in a respective range of formation combinations of between 25,000 and 29,000.

Firstly the data is broken down into individual days (roughly 235 flights each day), then all possible combinations (roughly 27k each time) are evaluated for formation. The unassigned formations dataset are then all the favourable formations from all combinations. These give a broader overview of all combinations and the spread of results. The MILP assigned formations, are then those which maximise the total possible saving. These are run for each day, and for different thresholds of alterations to takeoff time. The change in Take off time is important, as multiple identical flights on a day would quickly be put together otherwise, producing unrealistic results.
MILP assigned formations - with constraints
Using only the combinations whose take off times only need to be altered by a certain threshold

Unassigned formations
All combinations of flights evaluated for a given day