Verification of Chemical Plume Model for Use in Development of Autonomous Vehicle Tracking System

Abstract

A marine chemical spill environmental model that can accurately predict fine scale chemical concentration variations is needed in early stages of testing robotic control systems for tracking chemical spills. A combination Eulerian-Lagrangian type model, with two tuning parameters, was developed to produce the fine scale concentrations that would be observed by a robot; however, this type of model has not been compared to marine chemical spill data for evaluation or parameter tuning. Multiple model scenarios were run with different tuning parameters to determine the effects of those parameters on the model’s ability to reproduce an experimental measured chemical plume’s structure. A qualitative method for analyzing the concentration variations was established using amplitude and temporal statistical parameters. After tuning, the model was able to predict the real-time, fine scale chemical plume structure and can be used for testing control algorithms.