Determination and Prediction of Corrosion Integrity on RV Discovery Via Non-Destructive Test (NDT) And Artificial Neural Network (ANN)

Abstract

Marine environment is a harsh and severe environment for a metal structure like vessel, oil rigs and port infrastructure. A regular survey or monitoring is needed to reduce a structure failure due to the corrosion. To be seaworthy, a vessel should undergo a regular survey under specified timeframe. This survey is time consuming and costly. An alternative approach is required to predict the structural integrity of a vessel. Artificial Neural Network is one of the current methods that can be used to predict the deterioration rate of a structure. Corrosion integrity of RV-Discovery was determined via plate thickness measurement, coating thickness measurement and potential measurement. The data obtained from these measurements were used in artificial neural network to predict the deterioration rate. The results indicate that the plate and coating thickness reduction percentage is within minimal range while the average potential changes show that the structure is in passivation state. It implies that the structural integrity is in a good state with no or minimal maintenance required. The prediction of deterioration rate also shows that Scaled Conjugate Gradient (SCG) training algorithm was able to predict with over 95% of confidence and low mean square error.