Statistical Estimate of Determining the Critical Temperature of Brittleness for Metal of the VVER-1000 Reactor Vessel Using Impact Bending Test Data

Abstract

Using statistical modeling (Monte Carlo method), numerical experiments were performed to determine the critical temperature T_c of brittleness according to the PNAE G-7-002-86 and RD EO 0598-2004 methodologies. The data of the Charpy impact test (V-notched) used in the calculations were obtained using more than 1200 samples of 15Kh2NMFAA steel cut from various zones along the thickness, height, and circumferential direction of the VVER-1000 reactor vessel. The tests were carried out in the temperature range from –95 to +20°C. On the basis of statistical criteria, we show that the shell material of the reactor vessel can be considered as homogeneous. The values of the destruction energy of impact samples in the brittle-viscous transition region are found to be distributed according to a bimodal law. The distribution parameters are determined at various temperatures. Using the statistical modeling, the distribution laws of the critical temperature of brittleness are determined. The average values of T_c obtained using PNAE G-7-002-86 are shown to be approximately 10°C higher than those obtained using RD EO 0598-2004. We determined the boundaries of the intervals in which the values of the critical temperature of brittleness with 90% probability fall depending on the number of samples tested and the test scheme. Recommendations for improving the methodology of determining the critical temperature of brittleness are given.