Mechanistic verification of empirical UO2 fuel fracture models

• Published in: Journal of nuclear materials 2021-12, Vol.556 (C), p.153163, Article 153163
• Main Authors: Gamble, K.A., Knight, T.W., Roberts, E., Hales, J.D., Spencer, B.W.
• Format: Article
• Language: English
• Subjects: BISON; Computational grids; Computer applications; Confidence intervals; Cracking (fracturing); Empirical analysis; Finite element method; Fuel Fracture; Fuels; Irradiation; Light water reactors; Mathematical models; Maximum power; Nuclear fuels; Radiation; Randomization; Sensitivity analysis; Temperature gradients; Tensile strength; Uncertainty Quantification; Uranium dioxide; Uranium silicide; XFEM
• ISSN: 0022-3115; 1873-4820
• DOI: 10.1016/j.jnucmat.2021.153163

•The tensile strength of UO2 is the most sensitive material property determining crack formation.•Uniform and volume-weighted Weibull approaches to strength randomization result in similar fracture behavior.•A framework has been established to develop fracture correlations for advanced fuel types.•The uncertainty in the mechanistic fracture model envelope the predictions of existing empirical correlations. Standard UO2 fuel pellets used in light-water reactors fracture during irradiation due to the large thermal gradient in the radial direction. Over the decades, numerous researchers have explored fuel cracking from experimental and modeling points of view. To date, there have been both empirical and mechanistic approaches to predict the number of fragments that form in UO2. The empirical models only consider maximum power and burnup as inputs. Existing mechanistic approaches for normal operation have not accounted for irradiation effects. This work employs a mechanistic fuel cracking model using the extended finite element method to explore radial crack formation while including a sensitivity analysis that accounts for the randomization of tensile strength within the fuel, the strength randomization criteria (uniform or volume-weighted Weibull), power ramping rates, computational mesh density, maximum power level, and irradiation (burnup) effects. The results indicate that the uncertainty in this mechanistic modeling approach envelopes the predicted values from three different empirical correlations in almost all cases. This means that, for computationally intensive analyses involving UO2 fragmentation, the empirical correlations can be used. However, since the mechanistic calculations bound those of the empirical correlations, there is confidence in the applicability of the mechanistic approach developed in this work to generate a correlation for fuel types where limited data exists (e.g., doped-UO2, U3Si2).