Impact of ideal scaling on BEPU uncertainty and sensitivity analysis of the ACME test facility

In applications of the BEPU approach, uncertainty analyses are typically performed only at the test facility or the reactor, and the impact of model scale on uncertainty quantification remains an open issue to be further investigated. As a first step toward addressing this issue, an ideal scale-up comparative uncertainty analysis scheme is proposed, in which models of different scales for the target test facility are evaluated under identical accident scenarios using the same sampling set. Firstly, the realistically constructed ACME integral test facility is taken as the reference, and its numerical model is rigorously scaled up to the prototype plant size in strict accordance with scaling laws. The nodalization strategy, physical models, boundary conditions, and accident scenarios are also consistently preserved across scales. Then, both models are subjected to the same set of uncertain input parameters and an identical sampling matrix, and a representative 2-inch cold-leg-break loss of coolant accident transient is simulated for direct multiscale comparison. The resulting distributions of key safety-relevant responses are compared in terms of central tendencies, uncertainty bands, and consistency of safety margins, thereby directly quantifying scale-induced effects. The analysis shows that the multiscale BEPU results exhibit good coherence, scale-related biases remain within acceptable limits, and the extrapolation from test facility to prototype is technically justified.