Efficient global sensitivity analysis for flow-induced vibration of a nuclear reactor assembly using Kriging surrogates

Abstract

The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that: • a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.

Efficient global sensitivity analysis of structural vibration for a nuclear reactor system subject to nonstationary loading

“…Figure 9 shows the full-order SFEM GSA first-order sensitivities using FAST with 5,000 and 8,000 samples from which it can be seen that the difference in sensitivities upon changing sample size is negligible, indicating convergence of the sensitivity indices. Unlike the GSA results presented in (15) in which three of the six parameters imparted nearly zero variance to the model output of interest, five of the six parameters included in the SFEM GSA contribute at least 5% variance to the model output of interest (P17, peak lateral acceleration of the RVCH). Referring to Table 1 for the parameter identifiers, it may be seen that stiffness between two components in the reactor system 9 (P15) carries the greatest sensitivity, following by the number of components, 6 (P12).…”

“…This work demonstrates the extensibility of the surrogate-based GSA methodology shown useful for a reactor subassembly subjected to stationary random vibration in (15) to a reactor equipment SFEM, for which the finite element model included non-linearity and was subjected to non-stationary loading. The SFEM was parameterized with a total of sixteen model parameters for which there are plant-toplant variations or for which the magnitude changes over the course of the reactor life due to aging.…”

“…Furthermore, several works have already applied surrogate models for sensitivity analysis, successfully reducing the computational expense significantly such as (12), (13), (14). Specific to the nuclear industry, (15) This paper presents the application of surrogate modeling and GSA methods to non-stationary stochastic dynamic finite element analysis of a reactor system subject to a LOCA scenario. SFEMs are used ubiquitously throughout the design and analysis of key systems, structures, and components within pressurized water reactors (PWRs) across the commercial nuclear industry, per (16), (17) and (18), for example.…”

“…Despite the centrality of SFEMs to the general mechanical and civil design of such highconsequence facilities, UQ for such computational models is generally lacking amongst nuclear industry practitioners. As such, this work builds upon that of (19) and (15) by considering a non-linear finite element model with multiple outputs, additional input parameters, and non-stationary loading. In the following section, the methodology is described including the manner in which non-stationary random vibration was simulated using finite element analysis, the Fourier Amplitude Sensitivity Test (FAST) used to accomplish the GSA, Gaussian Process surrogate modeling, and then the design of experiment sampling strategies employed herein.…”

Effects of variable parameters on the behaviour of the single flexibly-mounted rod in a closely-packed array

“…In order to predict the typical behaviour of the system by using a low-dimensional representation for a high-dimensional problem, a mathematical approximation was applied. In the last years, model reduction methods are becoming more and more topical in flow-induced vibration research and analysis, for example, [19,20]. Response Surface Method is used to build approximation models in this study.…”