This paper on "The Use of Genetic and Amoeba Algorithms in Combination with a Lagrangian Solver to Simulate Dispersion and Deposition of Droplets Originating in Aircraft Wakes" was presented at the NAFEMS World Congress on The Evolution of Product Simulation From Established Methods to Virtual Testing & Prototyping - 24-28 April 2001, The Grand Hotel, Lake Como, Italy.

Summary

A widely used Lagrangian solver that calculates the trajectory and deposition of droplets released into the wake of an aircraft (Bilanin et al., 1989, Teske et a1. 1993) has been used as the computational module in a genetic algorithm to optimize serial spraying practice, and in an amoeba calculation to address deposition of fire retardant by aircraft. These approaches are invoked in two physical problems where large data sets have recently been collected This paper reviews both back calculation techniques, and discusses their implementation and results. The Lagrangian solver (AGDISP) uses inputs such as aircraft weight, wing semispan, spray material volatility, meteorology etc., to calculate the deposition of pesticides from aerial spraying. A recent study produced 185 model validation data sets (Bird et al., 01) and confirmed the model predictions are reasonable. The model can consider over twenty variables in the deposition calculation. In practice, it is often of more interest to specify the desired deposition rate and then investigate the mechanical system or environmental conditions that will allow that to be achieved. Unfortunately, back calculation through the model leads to combinatorial explosion (Teske et al., 1997). However, through the use of a genetic algorithm (essentially invoking repeated AGDISP calculations), a desired result (mass deposition per area) can be specified and the algorithm can be used to optimize operational variables such as release height or nozzle type (Potter et al. 00a, Potter et al. 00b). Similarly, an amoeba calculation can perform a statistical fit between ground deposition data and model prediction (again, with repeated AGDISP calculations), by optimizing four of the primary variables influencing the deposition of aerially released fire retardant material. This approach has been developed using a large database comprised of 800 bucket and tanker drops. The calculation backs out the best release conditions for each deposition pattern, and may be used to correlate similar bucket and tanker types, in a effort to develop a tool for predicting deposition in an untested drop configuration.