The calculation of the essential core safety parameters in fast spectrum reactors are loaded with large uncertainties due to the particularly large uncertainty of nuclear data, modelling uncertainties and possible mistakes by users. These core safety parameters such as coefficients of reactivity, power peaking factors and point-kinetics parameters play an important role during unprotected transients. The contributions to the uncertainties of the target parameters (e.g. peak cladding temperature) from the core safety parameters can be determined by uncertainty and sensitivity analyses.
Three-dimensional models of the ALLEGRO (demonstrator gas-cooled fast reactor) and a sodium-cooled fast reactor (SFR) concept were created in the Serpent Monte Carlo code, and several coefficients of reactivity were calculated. Furthermore, thermal hydraulic analyses with point kinetics were made for the selected transients by the ATHLET code system. An unprotected loss of flow and an unprotected overpower transient were studied at first using the best estimate approach to understand the behaviour of the reactors in these scenarios. In addition, uncertainty analyses were performed applying the GRS method, considering the uncertainties of the coefficients of reactivity, which were taken from benchmark results. Finally, the importance of the uncertain parameters was determined from sensitivity measurement.
The results show that the thermal expansion of the core structural elements has a significant effect on the reactivity of these fast spectrum reactors. Moreover, it was found that during unprotected overpower transients the most important uncertain parameters are the Doppler coefficient and the fuel thermal expansion coefficient of reactivity. In addition, the large uncertainty of the positive coolant and cladding temperature coefficients of reactivity have a significant influence on the uncertainty of the calculated peak cladding temperature, especially in the case of unprotected loss of flow transients. The results imply that the uncertainty of some core safety parameters should be reduced. Lastly, the results adumbrate that additional safety shutdown systems are to be applied to mitigate the consequences of unprotected transients.