Predicting Aerially Delivered Retardant Ground Deposit Concentrations and Spatial Distribution Using Statistical and Algebraic Modelling with Influence from Experimental Techniques

Date of Award

2022

Degree Name

Ph.D. in Aerospace Engineering

Department

Department of Mechanical and Aerospace Engineering

Advisor/Chair

Aaron Altman

Abstract

Combinations of various trends in global weather point towards an increased severity and frequency of wildfires. A handful of attempts have been made in the past that try to determine retardant ground deposits and their spatial distribution resulting from aerial drops in an effort to curtail fire growth. This study takes a multipronged approach at determining retardant ground deposits and spatial distribution at various coverage levels to better achieve fire control and extinguishment. The first approach fuses the dependent parameters, (line length, width, area, and ground distribution of the retardant), with the independent parameters using statistical regression in hopes to identify the probable parameters that are complicit in affecting the ground contours and their prediction the most. While the coverage prediction for the lower coverage levels (up to 3 GPC - Gallons per 100 ft.2) is accurate to 85% for area prediction with a variability of ±15% from actual while the length prediction is only accurate 58% of the time. This value was obtained using volume bounds on the input conditions. The estimate at higher coverage levels was poor along with the retardant's spatial distribution. An alternate approach was to model the drop phenomena in a relatively controlled, scaled down environment which was performed in the University of Dayton's Low Speed Wind Tunnel (UD - LSWT) facility. A 1 mm circular jet of water emanating from the underbelly of a model fuselage was placed in varying velocity crossflow (0 < Weber number < 101) of air. Shadowgraphs were initially performed, and the jet breakup was captured at 2000 frames per second which aided in discovery of breakup location with respect to surface waves. Experiments with Background Oriented Schlieren and Particle Image Velocimetry were also planned, however, they ultimately were not successful. Historical data points to two instabilities that govern the breakup process in jets, either in crossflow or quiescent air: the Kelvin-Helmholtz (KH) instability and the Rayleigh-Taylor (RT) instability. A code utilizing both the KH and RT instabilities to predict the atomization rates of the retardant as it issues from the aircraft and subsequently the retardant's ground coverage and its spatial distribution at the higher coverage levels. The results were compiled for 3 aircraft, namely the Aeroflite RJ-85, Minden BAe 146, and Neptune BAe 146 v2. The code had a prediction accuracy of 78% (for any given volume) for the 0.5 GPC coverage level line length with a variability of ±15% from actual. The results also capture the spatial distribution of the retardant coverage levels relatively well: a necessity for making consecutive drops for line buildup procedure. Furthermore, the entire simulation takes between 200 - 250 sec per simulation on a high-end personal computer, (Core i7 7700 HQ 2.80 GHz, 16 GB DDR4 RAM). Results of the current dissertation should allow the firefighting crew to better rapidly predict the intact line length and retardant spatial distribution aiding in better real-time planning, better resource management, and perhaps a quicker and perhaps more reliable extinguishment of wildfires.

Keywords

Aerospace Engineering, Aerial firefighting, Retardant ground prediction, Atomization/Jet in crossflow, Kelvin Helmholtz and Rayleigh Taylor Instability Waves, Shadowgraph/ PIV/ Background Oriented Schlieren

Rights Statement

Copyright © 2022, author

Link to Full Text

Share

COinS