Linear Elastic Fracture Mechanics (LEFM) and Stress Intensity Factor (SIF) are applicable tools to evaluate growth rate of existing fatigue cracks in offshore tubular joints. This is because the plastic zone around the fatigue crack tips is satisfactorily small. Several approaches based on LEFM have been proposed in this regard. Each of them uses different methods for estimating Stress Intensity Modification Factor (Y). In this research two types of Artificial Neural Networks (ANN) are trained for predicting the Y factor: Radial Basis Function (RBF) and Multi Layer Perceptron (MLP) networks. These networks are capable of estimating the Y factor at deepest point of a fatigue crack, located in tubular T joints under axial loading, when the crack depth is more than 20% of chord wall thickness. The required input data consist of the crack shape and the percent of the crack growth through thickness. Experimental data from NDE center in University College London are used for training and testing the networks. This data has been gathered by applying constant amplitude axial load to the brace of six full-scaled T joints. The results of this research are compared with different empirical and semi-empirical solutions, used in previous research works. The Y factors obtained from networks show better agreement with actual ones when compared to those obtained from previously established methods.

This content is only available via PDF.
You do not currently have access to this content.