Reactor protection system design alternatives for sodium fast reactors

Abstract

Historically, unprotected transients have been viewed as design basis events that can significantly challenge sodium-cooled fast reactors. The perceived potential consequences of a severe unprotected transient in a sodium-cooled fast reactor include an energetic core disruptive accident, vessel failure, and a large early release. These consequences can be avoided if unprotected transients are properly defended against, potentially improving the economics of sodium fast reactors. One way to defend against such accidents is to include a highly reliable reactor protection system. The perceived undesirability of the consequences arising from an unprotected transient has led some sodium fast reactor designers to consider incorporating several design modifications to the reactor protection system, including: self-actuated shutdown systems, articulated control rods, and seismic anticipatory scram systems. This study investigates the performance of these systems in sodium fast reactors. To analyze the impact of these proposed design alternatives, a model to analyze plant performance that incorporates uncertainty analysis is developed using RELAP5-3D and the ABR-1000 as the reference design. The performance of the proposed alternatives is analyzed during unprotected loss of flow and unprotected transient overpower scenarios, each exacerbated by a loss of heat sink. The recently developed Technology Neutral Framework is used to contextually rate performance of the proposed alternatives. Ultimately, this thesis offers a methodology for a designer to analyze reactor protection system design efficacy. The principle results of this thesis suggest that when using the Technology Neutral Framework as a licensing framework for a sodium-cooled fast reactor, the two independent scram systems of the ABR- 1000's reactor protection system perform well enough to screen unprotected transients from the design basis. While a regulator may still require consideration of accidents involving the failure of the reactor protection system, these events will not drive the design of the system. However, self-actuated shutdown systems may be called for to diversify the reactor protection system. Of these, the Curie point latch marginally reduces the conditional cladding damage probability for metal cores because of their rapid inherent feedback effects, but is more effective for the more sluggish oxide cores given reasonably long pump coastdown times. Flow levitated absorbers are highly effective at mitigating unprotected loss of flow events for both fuel types, but are limited in response during unprotected transient overpower events. When considered from a risk-informed perspective, a clear rationale and objective is needed to justify the inclusion of an additional feature such as self-actuated shutdown systems. The use of articulated safety rods as one of the diverse means of reactivity insertion and the implementation of an anticipatory seismic scram system may be the most cost-effective alternatives to provide defense in depth in light of the sodium fast reactor's susceptibility to seismic events.