Fusing Navigational and Meteorological Data for Prediction of Hull Slamming in Navy Patrol Boats (honours supervision)
Honours thesis by Katrina Kelleher (supervised by Stephen McAteer), presented by Katrina at MODSIM 2015.
Abstract: Driving a boat in high seas causes hull strain, which in turn can cause fatigue and costly damage. This paper investigates the feasibility of using weather data to predict hull strain measurements (in the form of slam counts) in the Royal Australian Navy fleet of Armidale Class Patrol Boats (ACPBs). Weather data containing the values of various wave attributes across the world’s oceans is combined with patrol boat GPS location data to develop a weather experience profile for a given ship.
This study uses interpolation, linear regression and big data techniques such as machine learning to investigate the interrelation of the datasets. A reliable correlation between the ship speed, wave frequency and slam counts is found and the null hypothesis rejected at the 99.9% confidence level. Several strategies for slam prediction are developed and tested however prediction accuracy is limited by insufficient data.
The approach discussed in this paper would allow maintenance inefficiencies to be reduced as a ship would potentially only require manual inspection when its weather experience profile indicates it is necessary. The ACPB fleet has an approximate whole of life cost of $650 million so finding efficiencies in the maintenance cycle has the potential to save the RAN millions of dollars.
(Australian National University, Katrina was awarded the HA Jones Medal for Excellence in Engineering Studies and Best Student Paper at MODSIM 2015).