Corrosion prediction for bulk carrier via data fusion of survey and experimental measurements

Z Wang,A.J Sobey,Y Wang

doi:10.1016/j.matdes.2021.109910

Abstract

Accurate corrosion predictions are vital to safe and optimised designs of marine assets. Traditional approaches, including those used to develop rule requirements, seek to use empirical regressions to model corrosion, but most are solely time-dependent. This may lead to conservative damage estimates and hence heavy and inefficient ships. To provide more accurate predictions, this paper presents an interpretable machine learning algorithm based on data fusion of ship survey and experimental measurements. The corrosion behaviour in bulk carrier ballast tanks is interpreted through a sensitivity analysis which quantifies the relationships between operational/environmental factors and the corrosion rate. The prediction accuracy is improved by a minimum of 82% when compared to the two representative empirical models, with a mean absolute error down to 0.10 mm.

Highlights

Marine corrosion of carbon steels has been extensively investigated
The model developed by Paik and Kim [52] for bulk carrier ballast tanks used the thickness measurement data collected from periodical surveys and a regression line to represent the relationship between the corrosion depth/rate and ship age
This paper proposes an Artificial Neural Networks (ANNs)-based data fusion approach using ship survey thickness measurements (TM) and corrosion experimental data to provide tailored corrosion predictions for sweater ballast tanks and decks of bulk carriers

Summary

Introduction

Marine corrosion of carbon steels has been extensively investigated. to accurately predict the corrosion rate of oceangoing ships, more than 90% of which are made of carbon steels, remains one of the most challenging tasks in materials science. The model developed by Paik and Kim [52] for bulk carrier ballast tanks used the thickness measurement data collected from periodical surveys and a regression line to represent the relationship between the corrosion depth/rate and ship age. This paper proposes an ANN-based data fusion approach using ship survey thickness measurements (TM) and corrosion experimental data to provide tailored corrosion predictions for sweater ballast tanks and decks of bulk carriers. The goal is to capture a wider range of influencing factors (both operational and environmental), as highlighted in bold, in the corrosion prediction process and to investigate the resulting method approximation of the fundamental understanding of marine corrosion mechanisms. The prediction was compared with current design requirements of corrosion allowances

Methodology to predict corrosion from a fusion of datasets

ANN with data fusion of ship survey TM and the experimental datasets

Predictions based on ship survey TM data

Predictions from data fusion of ship survey TM and experimental data sets

Sensitivity analysis

Conclusions