Abstract:This work illustrates a quantitative uncertainty analysis of CTF, one thermal-hydraulics sub-channel code for nuclear engineering applications. Like many other thermal-hydraulics codes, large uncertainties of CTF stem from the inaccuracy of parameters in the empirical correlations. However, many previous uncertainty analyses of CTF are based on the inexact parameter distributions described by the “expert opinions”, which are generally overestimations. To resolve these issues, inverse uncertainty quantification (IUQ) is conducted to determine the parameter uncertainties. IUQ is a process of quantifying the uncertainties of physical model parameters given relevant experimental measurements. In this study, the OECD/NRC BWR Full-size Fine-Mesh Bundle Tests (BFBT) benchmark steady-state void fraction data is used as the prior knowledge and Bayesian inference is implemented to formulate the IUQ problem. Two Markov Chain Monte Carlo (MCMC) sampling techniques, the adaptive Metropolis-Hasting with global scaling algorithm and the adaptive Metropolis-Hasting with component-wise scaling algorithm, are implemented and investigated to explore the posterior probability density functions. Once the framework is established, approximate 100000 runs of CTF are needed which is a big computational burden. Gaussian process regression (GPR) is therefore introduced to construct the surrogate model for CTF, which can significantly reduce the simulation time and make the Bayesian inference possible. The outcomes contain the posterior distributions of 7 physical model parameters that are crucial to the BFBT experiment. Results of IUQ can be used for the forward uncertainty propagation and validation analysis. The results also show that two MCMC algorithms take the advantages of self-adaptive method, capable of inferring the posterior distribution of model parameters with arbitrary form of prior information. These two MCMC algorithms give consistent results and the posterior accepted samples of each model parameter are then fitted to the classic distribution.

A kernel estimate method for characteristic function-based uncertainty importance measure

A fast computational method for moment-independent uncertainty importance measure.

A stable approach based on asymptotic space integration for moment-independent uncertainty importance measure.

A new moment-independent uncertainty importance measure based on cumulative residual entropy for developing uncertainty reduction strategies

Computational strategy for uncertainty importance measure ranking based on norm

Unified Reliability Measure Method Considering Uncertainties of Input Variables and Their Distribution Parameters

An Estimate of a Change Point in Variance of Measurement Errors and Its Convergence Rate

A method of measuring the difference between probability models as uncertainty and its applications in financial economics

Efficient Input Uncertainty Quantification for Ratio Estimator

Quantify uncertainty by estimating the probability density function of the output of interest using MLMC based Bayes method

Measures of Uncertainty Based on Gaussian Kernel for Type-2 Fuzzy Information Systems

Non‐parametric measure approximations for constrained multi‐objective optimisation under uncertainty

Efficient method for fully quantifying the uncertainty of failure probability

The Importance Measure of Fuzzy Input on Failure Credibility under the Fuzzy Uncertainty

A new efficient simulation method based on Bayes' theorem and importance sampling Markov chain simulation to estimate the failure-probability-based global sensitivity measure

Uncertainty Quantification and Confidence Intervals for Naive Rare-Event Estimators

INVERSE UNCERTAINTY QUANTIFICATION OF CTF PHYSICAL MODEL PARAMETERS USING BAYESIAN INFERENCE

Uncertainty Characteristics Curves: A Systematic Assessment of Prediction Intervals

An efficient method based on Bayes’ theorem to estimate the failure-probability-based sensitivity measure

EFFICIENT RISK MEASURES CALCULATIONS FOR GENERALIZED CREDITRISK+ MODELS

Kernel estimation of multivariate cumulative distribution function