A probabilistic modelling of corrosion growth in marine ballast tank for sustainable maintenance scheme

Abstract

This paper exhibits the application of probability distributions in modelling the corrosion growth behaviour of pits in vessel's marine ballast tanks. The metal loss data gained from real inspection has been carried out to model the defect growth. However, the provided information from the vessel inspection is full of uncertainties owing to the nature of marine corrosion. Hence, the time-dependent Weibull model is proposed in order to cater the variation of parameters and is capable to predict the distribution of corrosion depth at any point of vessel's age by eliminating environmental factors, material properties and operational condition. The predicted data as compared to actual data yields results with moderate accuracy based on Root-Mean-Square-Error (RMSE), yet still promising provided that more high quality data can become available in the future. The proposed model intends to simplify the modelling process so the available data can be fully utilised for future prediction purposes, hence improving the traditional prescriptive inspection planning of corroding structures. If more information can become available, the prediction model can be improved to achieve a higher accuracy of depth prediction in the future.