Nuclear Technology / Volume 166 / Number 3 / June 2009 / Pages 204-214
Technical Paper / 2007 Space Nuclear Conference / Fission Reactors
This paper focuses on some of the unique dynamic characteristics of compact fast-spectrum reactors. The study is limited to the characteristics that are relatively independent of how the reactor is integrated into a complete power system. Some of the well-established characteristics of compact fast-spectrum reactors are that point kinetics is generally very accurate for these systems and that temperature and burnup reactivity feedback mechanisms are relatively small and simple. Beyond this, there are two unique aspects of highly reflected fast reactors (e.g., space reactors) that do not occur in more traditional reactors. First, the neutron reflector has a very important impact on dynamic performance, and in some cases the temperature coefficient of the radial reflector is higher than that of the fuel. The thermal time constant of the reflector is much longer than that of any component in the core, which requires all reflector temperature and expansion effects to be modeled individually. Second, reflected neutrons have a much longer fission life span than in-core neutrons. In effect, this creates additional delayed neutron groups, referred to as geometric delayed neutron groups. These groups can have life spans orders of magnitude longer than neutrons that do not leave the core, and have much higher worth due to moderation. For compact beryllium reflected reactors there is also a measurable delayed group of photo-induced neutrons that result from delayed gammas. Another characteristic of compact fast-spectrum reactors is simplified control and the ability to passively handle a wide range of transients without control. Various transient analyses are presented that were performed by the Fission Reactor Integrated Nuclear Kinetics (FRINK) code, which facilitates near-term compact reactor design and development by providing a transient analysis tool. In its current state FRINK is a very simple system model, and the "system" only extends to the primary loop power removal boundary condition; however, this allows the simulation of simplified transients (e.g., loss of primary heat sink, loss of flow, etc.).