Programmer's Guide to UnMES Construction and Spatial Application Documentation for UnMES Transition Package to NRL-SSC Final Report- Part B
Abstract
This document describes the software required to build, train and use the Underwater Munitions Expert System (UnMES), a probabilistic management tool designed to predict underwater migration and burial of abandoned munitions. UnMES is based on a Bayesian Network framework that relates the geological and hydrodynamic site characteristics to the processes governing munitions' interaction with sediments, waves and currents. Simple deterministic models of causal forces acting on the underwater munitions and the sediment responses have been developed to predict scour burial, onset of motion, and the potential for burial or re-exposure due to bedform migration [Rennie, Brandt and Ligo, 2019, hereafter RBL2019]. These models were implemented in Matlab [Mathworks, 2022] at JHU/APL. Other preliminary models for the estimation of liquefaction burial and total distance migrated are presented, despite lack of full validation. These deterministic models are used to build a data set to train the conditional probability tables forming the core of UnMES.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 31, 2022
- Accession Number
- AD1187739
Entities
People
- Sarah Rennie
Organizations
- Johns Hopkins University