An artificial neural network modeling for pipeline corrosion growth prediction

Abstract

Corrosion defect assessment becoming a forte issue in pipeline reliability assessment to accurately predict the severity of its condition. Due to the uncertainties inherit from the pipeline inspection at present, statistical model use to model the corrosion growth apply a correctional methods to reduce the gap (means and variation) between predicted values and the actual data. This study aims to develop a time dependent corrosion growth model for oil and gas pipeline using Artificial Neural Network (ANN) as an alternative to the current method and to evaluate its applicability without enforcing data correctional methods. This model is formulated based on parameters of defect extracted from in-line inspection data (ILI) and quantified by statistical analysis. The develop model gives the prediction of the corrosion depth and length of the defect that can be used to calculate the corrosion rate or growth. The results and outcome of the present study can help pipeline operators to predict the reliability of the pipeline structure in terms of its probability of failure or lifetime estimation