Probabilistic Estimation Plastic Properties of Alloy 600 From Spherical Indentation Using Surrogate‐Assisted Bayesian Inference Inverse Approach

Alloy 600, a nickel‐based material, is widely used in steam generator tubing and penetration nozzles of pressurized water reactors (PWRs) due to its exceptional corrosion resistance and excellent mechanical properties under high‐temperature and high‐pressure water conditions. However, conventional uniaxial testing, which is commonly employed to obtain plastic parameters, exhibits significant limitations in practical structural assessments of Alloy 600 due to its destructive nature. To investigate the local mechanical properties of Alloy 600 while accounting for the uncertainty in material parameters during indentation testing, this study proposes a probabilistic inverse method combining Bayesian inference with spherical indentation testing. To enhance the computational efficiency of the likelihood function in Bayesian theory, a preprocessing and postprocessing program for finite element (FE) software was developed to enable batch computations. Furthermore, an adaptive PC‐Kriging surrogate model was employed to replace FE simulations, achieving efficient approximation of high computational cost models. Finally, Markov Chain Monte Carlo (MCMC) sampling was systematically applied to analyze the probabilistic distribution characteristics of Alloy 600’s plastic properties. Comparisons with traditional uniaxial tensile tests and other Bayesian methods confirm the validity and reliability of the proposed approach. The results demonstrate that this method provides probabilistic estimations of plastic properties, including probability distributions and confidence intervals, with stress‐strain curves showing strong consistency with uniaxial tensile test data. The novel approach exhibits remarkable effectiveness in the probabilistic characterization of Alloy 600’s plastic properties.