ANALYSIS OF 3D SEMI-ELLIPTICAL CRACK ON REACTOR PRESSURE VESSEL WALL WITH LOAD STRESS AND CRACK RATIO

Reactor Pressure Vessel (RPV) wall is an important component in the Nuclear Power Plant (NPP). During reactor operation, RPV is subjected to high temperature, pressure, and neutron exposure. This condition could lead to RPV structure failure. In order to assure the integrity of RPV during the reactor lifetime, it is mandatory to perform a structural integrity assessment of RPV by evaluating postulated crack in RPV. In the previous study, the crack has evaluated in 2-D. However, 3-D analysis of semi-elliptic crack shape in the surface of the thick plate for RPV wall using SA 508 Steel is yet to be analyzed. The objective of this study is to analyze and modeling the evaluation in variation crack ratio with some load stress in 3-D. The Stress Intensity Factor (SIF) and J-integral are used as crack parameter. The J-Integral were calculated using MSC MARC MENTAT based on Finite Element Method (FEM) for obtaining the SIF value. The inputs are a crack ratio, load stress, material property, and geometry. The modeling of SIF value and goodness of fit are using MINITAB. The fracture condition could be predicted in comparison to the SIF value and fracture toughness. For the load stress 70 MPa and 80 MPa, with a crack ratio 0.25, 0.33 and 0.5, the material on RPV wall will in fracture condition.Keywords: Semi elliptic surface crack, 3-dimension, reactor pressure vessel, elastic-plastic fracture mechanics, J-integral

Comparing Different Metamodelling Approaches to Predict Stress Intensity Factor of a Semi-Elliptic Crack

This paper examines the applicability of the different meta-models (MMs) to predict the Stress Intensity Factor (SIF) of a semi-elliptic crack propagating in topside piping, as an inexpensive alternative to the Finite Element Methods (FEM). Five different MMs, namely, multi-linear regression (MLR), second order polynomial regression (PR-2) (with interaction), Gaussian process regression (GPR), neural networks (NN) and support vector regression (SVR) have been tested. Seventy data points (SIF values obtained by FEM) are used to train the aforementioned MMs, while thirty data points are used as the testing points. In order to compare the accuracy of the MMs, four metrics, namely, Root Mean Square Error (RMSE), Average Absolute Error (AAE), Maximum Absolute Error (AAE), and Coefficient of Determination (R2) are used. Although PR-2 emerged as the best fit, GPR was selected as the best MM for SIF determination due to its capability of calculating the uncertainty related to the prediction values. The aforementioned uncertainty representation is quite valuable, as it is used to adaptively train the GPR model, which further improves its prediction accuracy.