Statistical Model for Prediction of Plate-Specific Fracture Toughness Properties of ASTM A285 Steel

Abstract

A materials test program was developed to measure mechanical properties of A285 carbon steel under conditions relevant to waste storage tanks at the Savannah River Site. Fracture toughness and tensile testing were performed on ASTM Type A285 steels that span tank plate compositions. Variables relevant to the material and load conditions for Type I and Type II tanks were defined and a statistical test matrix was designed for fracture toughness testing. The test matrix consisted of eight variables expected to influence mechanical properties. The independent variables were carbon content, manganese content, sulfur content, thickness, pearlite fraction, and grain size. The responses were the fracture toughness JIc, and a figure-of-merit at J3mm to allow for sub-critical ductile crack growth. A total of 85 J-R curves were compiled. The full data set was used to construct statistical models to predict fracture properties as a function of metallurgical properties and operating conditions. Eight independent models resulting from combinations of loading rate (quasi-static, dynamic) and orientation (T-L, L-T) were developed. The statistical significance of terms was determined for each of the models. Thickness and grain size were found to be of statistical significance in the models developed for the dynamic strain rate testing data. Compositional variables were found to be of statistical significance for the quasi-static loading rate fracture toughness data.