Author: Jesse Stewart
Stewart, Jesse, 2019 Design of an automated flight planning and fleet allocation tool that optimises the delivery of water from water sources, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Wildfires are becoming an increasingly common occurrence across Australia with both fire intensity and surface area impact for each fire event continuing to grow. As the frequency of wildfires increase and rural firefighting resources are strained the benefit brought by aerial fire suppression is becoming more critical for the timely containment of wildfires. The highly dynamic environment encountered during aerial suppression operations requires a system that can adapt to the changing conditions with minimal impact to operational efficiency. Available aircraft must be assigned to currently active fire fronts, have their flight paths optimised to minimise the distance travelled, service the fire fronts while meeting the objectives of the suppression efforts, and utilise all available water sources to minimise their turnaround time. The challenges faced during management of aerial suppression resources amount to a fleet allocation and flight planning problem.
Building upon analysis of prior work in the fields of wildfire containment strategies, fleet management, vehicle routing, constraint-based planning, and flight planning, a fleet allocation and flighting planning system is proposed. The proposed system uses a layered approach allowing for a series of task specific algorithms to be integrated to form a unified solution. The fleet allocation layer implements a fire characterisation protocol to inform aircraft assignments to active fire fronts. The flight cost approximation layer generates an approximate flight cost by accounting for airspace to avoid built-up areas, represented by cost regions. Finally, the flight planning layer adopts an augmented greedy algorithm for the initial assignment of aircraft followed by an analysis of the expensive flight assignments, optimising where possible through assignment swapping.
The fleet allocation and flight planning system developed was implemented behind a prototype user interface to demonstrate the systems capabilities. The user interface implemented consists of a map as the primary interface window allowing for entities such as water sources, airbases, and fires to be introduced into the system through a mouse click. In addition to the testing and demonstration capability it affords, the user interface provides a unique and efficient tool for the generation of test data. The resulting system provides a reliable means of allocating available aircraft following an assessment of the fire, generating an approximate flight cost likely to be incurred, and forming a flight plan for the distribution of aircraft to fires and water sources. Incorporated alongside, where possible, the system provides a means of strategically utilising short-term resources to maximise the efficiency of aerial suppression resources.
Ultimately, this project was able to identify through both an extensive research phase and vigorous development, the key components of an intelligent fleet allocation and flight planning system. Using the insights gained, a novel system was implemented for allocation and planning of aerial suppression aircraft. The solution was designed such that it could be utilised separate from the aerial firefighting example case and is domain independent. Deliberate decisions made throughout the development process has afforded the system the potential to form the basis for future intelligent fleet management applications.
Keywords: Path planning, fleet allocation, flight planning, aerial firefighting, wildfires, bush fires, firebombing, intelligent fleet management, automated
Subject: Engineering thesis
Thesis type: Masters
Completed: 2019
School: College of Science and Engineering
Supervisor: Karl Sammut