Prediction of Mechanical Properties of Sensitized Stainless Steel by Neural Network Modeling and Validation Using Ball Indentation Test

Abstract

Elevated temperature sensitization of a 304 stainless steel results in degradation of mechanical properties and becomes prone to premature failure. In the present investigation, sensitization of 304 stainless steel has been done in the temperature range of 500–800 C. Yield strength, ultimate tensile strength and fracture toughness (KJc) of the sensitized 304 stainless steel specimens were determined by ball indentation technique. Microstructural characteristics were quantified and used in artificial neural network to predict the mechanical properties of the investigated alloy. Neural network was developed with the help of MATLAB toolbox. Best equation was fitted for training, testing and validating the output. Predicted values from the developed model exhibited impressive correlation with experimental data obtained through ball indentation technique as well as with literature reports. The model has proved its distinctive potential in predicting the mechanical properties of sensitized 304 stainless steel, which faces restriction in bulk sampling from original component to perform conventional mechanical test during service exposure.