Self-Organisation for Wildfire Supression

Lead Researcher: Dr Mauro S. Innocente
Researcher: Paolo Grasso

1)  Self-Organising Swarm of Massless Firefighting Drones

• Mass-less drones self-organised based on Particle Swarm algorithm adapted to individuals operating within physical dynamic environments of high severity and frequency of change.
• Realistic Fire Propagation Model, though only enthalpy balance and chemical species balance are considered at this stage for efficiency.
• Includes combustion, 2D diffusion, pseudo-3D radiation, pseudo-3D convection (though no conservation of momentum).
• A given ignition temperature triggers a fire, which consumes “forest fuel”.
• Combustion does not take place if temperature is below ignition or “forest fuel” has been depleted.
• Fire propagation model solved using centred finite differences in space and the general Runge-Kutta method in time.
• When a drone releases a water payload drop, the nodal temperature decreases reaching a value that results from a mass-based weighted average of the temperatures of the payload and of the node.
• Drone docking stations located at the top left corner in the form of a small grid (one specific location assigned to each drone with a 1-metre gap between them).
• Water payload source located at the bottom left corner in the form of a small grid (one specific location assigned to each drone with a 1-metre gap between them). Locations of docking stations and water payload sources coincide.
• Battery consumption measured by distance travelled.
• Excluding visualisations, this autonomous firefighting system simulation runs faster than real-time using a standard PC (including fire propagation).

Example of initial swarm of 100 drones