The current safety basis (SAR-153) for the Advanced Test Reactor (ATR) ensures that the plant protection criteria is maintained for all Condition 2 events by verifying that, for a Condition 2 Flow Coastdown Transient and Condition 2 Reactivity Insertion Accident, the Departure from Nucleate Boiling (DNB) Ratio (DNBR) is greater than two. The basis used to establish this limit is not well defined but may be traced to research reactor licensing based on overly conservative thermal hydraulic criteria. This limitation may not be applicable to reactor experiments because the quantity of fissionable material and fission product inventory in experiments is much less than that of the reactor core, and may prevent or limit future experimental testing in the ATR. In particular, fueled experiments may be excluded from irradiation in ATR if the desired fission power cannot be achieved due to these safety criteria.

This study will evaluate the DNBR using various CHF correlations and consider the impacts of changing the limit to a more suitable thermal hydraulic safety limit for fueled experiments, such as Onset of Nucleate Boiling (ONB), Onset of Significant Voiding (OSV), and Onset of Flow Instability (OFI). The study utilizes a BEPU (best-estimate plus uncertainty) statistical approach that maintains 3? from thermal hydraulic safety limit during Condition 2 transients. Modeling parameters of different correlations will be evaluated by considering the associated uncertainties. The study will eventually provide recommendations to support any safety basis changes that can expand the experimental operating envelope of the ATR without a compromise in safety.

The preliminary results indicate that large safety margins can be kept for Critical Heat Flux (CHF) based DNB at almost all conditions. DNBR only reaches slightly below two when using Groeneveld’s CHF Look-up Table (LUT) at peak transient condition with all conservative power multipliers applied. At all studied conditions, OFI ratios are found always greater than two and significant margins have been kept from OSV. ONB only occurs at steady-state when all conservative power multipliers are applied. Preliminary results involving an additional method with Dakota/RAVEN coupled to RELAP5 is presented.