The availability of metal-fueled fast reactors can be improved by the elimination of scrams caused by a class of low-flow events by taking into account the inherent reactivity feedbacks acting during such events. With this approach, the flow signal normally used in the loss-of-flow (LOF) protection system has its trip setpoint substantially reduced, and the power-to-flow signal is used with a normal trip setpoint. This permits the safety system to more closely monitor the safety of the reactor, i.e., its temperature rise as indicated by the power-to-flow ratio, and to avoid the initiation of a scram during an event in which the flow is decreasing but the core temperatures are remaining within prescribed limits. During such events, the inherent feedbacks act to reduce power and thereby limit any temperature or power-to-flow increases despite substantial reductions in the flow. This approach was applied to such an incident in Experimental Breeder Reactor II and was shown to avoid the scram that actually occurred while providing reactor protection. Additional calculations were performed illustrating the range of applicability of this type of LOF protection.

A series of tests was conducted in EBR-II which examined the dynamics of the transition from forced to natural circulation flow in a liquid-metal-cooled fast breeder reactor. Each test was initiated by abruptly tripping an electromagnetic pump which supplies 5–6 percent of the normal full operational primary flow rate. The ensuing flow coastdown reached a minimum value after which the flow increased as natural circulation was established. The effects of secondary system flow through the intermediate heat exchanger and reactor decay power level on the minimum in-core flow rates and maximum in-core temperatures were examined.

An experimental and theoretical program has been undertaken during the past several years with the objective of developing a well-documented understanding of steady-state and transient thermal-hydraulic behavior in EBR-II. The results of this effort have provided reactor designers and system modelers with needed integral-type demonstrations of important phenomena. This paper will discuss the particular problems of steady-state and transient hot channel peaking factors and plant operational characteristics impact upon natural circulation dynamics. Direct in-core experimental measurements have demonstrated that factors used for the prediction of peak coolant temperature rises at normal rated plant conditions may not be conservative due to pin-bundle distortions or inlet flow maldistributions, while those applied during loss-of-flow transients are most likely overconservative due to inter- and intrasubassembly phenomena. The importance of somewhat controllable parameters such as the sequence of primary and secondary pump trips and reactor scram, primary pump rundown times, and nominal operational power-to-flow ratio upon the dynamics of the transition from forced to natural convective flow are also presented.

In the analysis of core-wide temperature distributions in a liquid-metal-cooled fast breeder reactor (LMFBR) under both normal and abnormal operating conditions, it is commonly assumed that there is little, if any, thermal interaction between adjacent subassemblies. Since intersubassembly heat transfer tends to reduce the transverse temperature gradients in a reactor, thereby ameliorating the effects of local overheating, this assumption is conservative. In order to assess the importance of this effect as well as of flow redistribution in a reactor core, an experimental study was conducted in EBR-II covering a wide range of operating power and flow conditions, including both forced and natural convection. The results of this study indicate that radial heat transfer and flow redistribution are important mechanisms in the thermal-hydraulics of LMFBR cores, especially at low flow rates.

Conservation equations are derived for the motion of a small inert gas bubble in a large flowing liquid-gas solution subjected to large thermal gradients. Terms which are of the second order of magnitude under less severe and steady-state conditions are retained, thus resulting in an expanded form of the Rayleigh equation. The bubble dynamics is a function of opposing mechanisms tending to increase or decrease bubble volume while being transported with the solution. Diffusion of inert gas between the bubble and the solution is one of the most important of these mechanisms included in the analysis. The analytical model is applied to an argon gas bubble flowing in a weak solution of argon gas in liquid sodium. Calculations are performed for these fluids under conditions typical of normal and abnormal operation of a liquid metal fast breeder reactor (LMFBR) core and the resulting bubble radius, internal gas pressure, and mass of inert gas are presented in each case. An important result obtained indicates that inert gas bubbles reaching the core inlet of an LMFBR will always grow as they traverse the core under normal and extreme abnormal conditions and that the rate of growth is quite small in all cases.

Measurements of the vapor growth patterns and rates following the nucleation of superheated sodium in a vertical rectangular channel are presented and discussed. The vapor was found to grow as a single bubble for incipient bulk-liquid superheats greater than about 10 deg C, and this single bubble tended to completely fill the channel cross section (except for a thin liquid film on the walls) and to grow as a vapor slug for incipient bulk-liquid superheats greater than about 50 deg C. The temperature gradients in the liquid both normal and parallel to the channel axis prior to nucleation were found to have an important effect upon the dynamics of the vapor slug. Experimental data on the vapor growth and collapse rates and the associated pressure transients are presented for boiling pressures up to 1 atm and incipient superheats up to about 180 deg C.

"An Experimental Demonstration of Inert Gas Effects upon the Incipient Pool Boiling of Sodium." Nuclear Science and Engineering, 42(3), pp. 427–428

An experimental study is presented of the transient and steady-state natural convection heat transfer in mercury following a step change in uniform heat flux of the vertical rectangular channel wall. The fluid was initially stagnant and at a uniform temperature. Temperature distributions were measured during the transient and at steady-state conditions. Heat transfer characteristics of the system were determined from these measurements.