Click for next page ( 59


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 58
6 Nuclear Criticality Considerations As discussed in this chapter, the pane! finds that the probability of a critical excursion during the processing of the Molten Salt Reactor Experiment (MSRE) salt is extremely low. Even if such an event were to occur, the safety and technical consequences would be insignificant. Additionally, certain process options (e.g., see Appendix D) can reduce the likelihood of criticality even further. However, the public concerns and political consequences could be very large; thus, the pane} has addressed the question in some detail. CRITICALITY ISSUES IN PROCESSING Nuclear criticality safety considerations provide significant restrictions to the process designs to be implemented for each of the processing options. Nuclear criticality safety issues are determined by the chemical and physical behavior of the constituents, for example, the potential for uranium hexafluoride (UFO) evolution and condensation, the lower uranium fluorides produced, uranium reduction to metal, and zone refining as a result of melting. Table 6.1 outlines some specific concerns for each remediation operation. A more general nuclear criticality issue that is not specific to any one processing option is the potential consequence of a nuclear criticality excursion. Two elements of relevance are the shutdown mechanism that would limit the consequences of such an excursion in the salt medium and the range of energy releases from such an event. Pruvost and Paxton (1996) describe initiating mechanisms and consequences for historical and other postulated nuclear criticality accidents. Augmented neutron flux, heat generation, and production of new fission products would be signs of an excursion were it to happen. If the MSRE salt is inhomogeneous, it is not evident that a mechanism exists that would vary the composition to concentrate 58

OCR for page 58
NUCLEAR CRITICALITY CONSIDERATIONS TABLE 6.1 Nuclear Criticality Safety Implications: A Sample of Concerns Considered by the Pane} 59 Remedial Operation General: Verify current subcritical configuration Criticality Hazard/Concern Baseline: Establish (best-estimate) critical mass values for potential compositions of 233u in dry salt in (unrestricted) aqueous solution with effects of water and other moderators Compensatory Measures Maintain configurations and compositions Monitor for changes Prevent additions of moderators Apply adequately conservative safety factors (e.g., based on assumptions that bound uncertainties in supporting data, which are known to be inadequate, arid calculations) 1. Remove reactive gases Ingress of water or other Strict control of moderator to drain tanks moderator sources Favorable geometry in NaF traps, alumina, and zeolite bed 2. Remove solid UFO Redistribute existing Apply batch/mass deposits material to critical limits configuration Accumulate external critical mass Ingress of water or other moderator Remove to favorable geometry containers Strict control of moderator sources

OCR for page 58
60 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE 3. Sample solid salt Ingress of water or other Strict control of moderator moderator sources Collapse of salt Mechanical reconfiguration of salt 4. Remove residual Ingress of water or other Strict control of uranium in piping moderator moderator sources 5. Fluorinate in place to UFO 6. Remove salt as liquid for fluorination elsewhere 7. Remove salt as solid Accumulate external critical mass Ingress of water or other moderator Redistribution or precipitation of uranium or plutonium Uranium redistribution, segregation, or precipitation to form critical mass Ingress of water or other moderator Ingress of water or other moderator Collapse of solid Remove to favorable geometry containers Strict control of moderator sources Fluorinate carefully with HE, followed by F2 Apply batch/mass limits Remove to favorable geometry containers Analyze and establish controls for new external systems Strict control of moderator sources Control geometry and/or mass of removed liquid material Strict control of moderator sources Salt removal procedures to stabilize geometry and prevent collapse

OCR for page 58
NUCLEAR CRITICALITY CONSIDERATIONS 8. Convert UFO to U3O8 Accumulate critical mass Ingress of water or other moderator Note: This conversion step is well established, having been conducted frequently at ORNL (for 235U-bearing materials as well as for U-bearing materials) over the last several decades 61 Strict control of sources of water and other moderators (or apply limits and/or geometries that accommodate optimum moderation) Apply batch/mass limits and/or use favorable geometry containers NOTE: F2 = molecular fluorine; HE = hydrogen fluoride; NaF = sodium fluoride; 0RNL = Oak Ridge National Laboratory; U = uranium; UFO = uranium hexafluoride; U3O~ = uranium oxide. fissile material at the rapid rate required to sustain a nuclear chain reaction (i.e., to go critical). Without a rapid assembly or concentration mechanism, even if a chain reaction is achieved the resulting energy release would be very small and a breech of the drain tank containment would not occur. Further, the configuration of the MSRE system includes both distance and heavy shielding, which were enough to provide protection during operations at a steady power level of ~ MW. The radiation effects of a potential critical excursion would be unlikely to spread beyond the enclosed vessels; evacuation of surrounding rooms and buildings would be unnecessary. Thus, the consequences of a criticality excursion within the drain tanks are manageable. The pane! concludes that, even if such an event were to occur, the safety and technical consequences would be insignificant. CRITICALITY HAZARD OF REMELTING THE FLUORIDE SALTS IN THE DRAIN TANKS Aside from these criticality issues for various process alternatives, an assessment can be provided on the low nuclear criticality hazard when remelting the salt. The salt mixture is less effective as a moderator than water. However, three of its four constituents are "good"

OCR for page 58
62 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE moderators. Specifically, the effectiveness of lithium and beryllium is explained because they are lower in atomic number than the carbon (graphite) used to provide extra moderation for the MSRE core during operation. Fluorine comes after carbon in the periodic table, but it is still of low enough mass to be considered a moderator. Zirconium has a relatively high mass, with reduced moderating effectiveness. The overall combination of the elemental salt constituents provides a somewhat less effective moderator than graphite, which is one reason graphite was added as a moderator to the MSRE core design (at some substantial cost and inconvenience). The composition and concentrations of the salt mixture are such that even in the optimum configuration, criticality could not be achieved in the MSRE reactor vessel without the addition of a graphite moderator. Without as good moderation as in the graphite core, and with reduced uranium content (each drain tank originally contained only approximately half the uranium inventory, and more than 10 percent of that has migrated out of the salt), the probability of a criticality excursion in each drain tank is further reduced. Additionally, the pane} notes that the shielding was clesigned to be adequate for sustained criticality during reactor operations. One approach to assessing criticality potential is to use model calculations that generate a keff value for a particular geometry of fissile i ~ there 233U 23su and 239Pu tplutonium-2393~; neutron- moderating material (here, water, lithium, beryllium, and fluorine), and reflecting material (here, salt and concrete). Presentations to the panel (Rushton et al., 1996a,b) have included such calculations. Criticality is possible for the case of water intrusion into the salt (interior to the tanks), and for this reason, moderator controls are important. However, these calculations have shown (i.e., have generated a ken value much less than one) that external moderation alone cannot cause criticality. The nominal uranium-in-salt concentration (i.e., approximately ~3 g of uranium per liter, if half of the original uranium content is assumed to be distributed equally among the two tanks) is insufficient to be critical in a drain tank. A local increase in concentration could lead to a critical configuration in a tank. Should the concentration increase by a factor of two in a large enough, sphere-like configuration, criticality might be possible. Conversely, if the current uranium concentration in the drain tank remains relatively uniform (or, as is possible, is well below the

OCR for page 58
NUCLEA R CRI TI CA LI TY CONSIDERS TI ONS 63 nominal value), a well-controlled melt is unlikely to lead to a critical configuration. RATIONALE FOR TECHNICAL INSIGNIFICANCE OF A CRITICALITY EXCURSION The pane} believes that a criticality excursion, even were it to occur, would have no important technical consequence. Any operation would be done by remote means; that is, no human would be inside the drain tank cell. The drain tank cell provides massive concrete shielding (several feet thick; see Figure 1.3) and a gas- tight steel liner. Any criticality excursion event would be totally contained and well shielded. The building ventilation system provides the necessary secondary containment. This containment system was designed for an operating reactor and thus should contain any plausible criticality excursion in the drain tanks. A comparison of the MSRE drain tank system to that of previous criticality accidents provides perspective. The latter involved mechanisms to concentrate fissile material, for example, by a continuous feed to a reaction. From a historical and practical point of view (Stratton and Smith, 1989; Frolov et al., 1995; Knief, 1985; Pruvost and Paxton, 1996) a configuration similar to the MSRE reactor vessel is required for this. Such a configuration is absent in the case of the MSRE drain tanks. For example, were the tank wall to leak or to rupture (as a worst-case scenario), the material would disperse inside the drain tank cell and be subcritical in a slab geometry configuration. The cleanup problem would be transferred from the tanks to the cell. Even the rupture of the half-inch-thick Hastelloy N drain tank wall (which, if it did occur, would still pose no significant risk to the public) is a conservative scenario for a criticality excursion, with the following rationale: any criticality burst could take place only if the salt were liquid (i.e., at a temperature greater than 460C), but beryllium fluoride tBeF2] sublimes at 800C (at one atmosphere). Therefore, a temperature rise of 340C, obtained from any fission burst, would automatically disassemble the salt system, just as it would in an aqueous system (the BeF2 gas void would be similar to the steam void in a water

OCR for page 58
64 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE system). In both systems the nuclear reaction automatically shuts down after an initial burst. Without a breach of vessel walls, a criticality event may be detectable only by the presence of fresh fission products. CONCLUDING COMMENTS The panel had neither the resources nor the charter to perform a quantitative risk assessment (of criticality or of any other hazard) and notes that a rigorous effort is premature until the specific event scenarios and process conditions are defined. Nevertheless, the basis for the panel's position is found in the approach suggested in Appendix E to evaluate the criticality probability and to limit it to less than lob per year. The likelihood of a significant dose to someone at the site boundary is several orders of magnitude lower, due to the small fission product inventory, the relatively low energy release (on the order of megawatt-seconds) that is hypothetically possible from criticality (as a worst case), and the presence of significant containment barriers. Detailed analyses of credible uranium and salt configurations are continuing at Oak Ridge National Laboratory (ORNL). Based on a review of previous nuclear criticality safety evaluations, the pane! has confidence in the ORNL analysis and evaluation capabilities. Further, it is assumed that with whatever processing option is employed, the associated nuclear criticality safety measures will include (1) requirements that all activities proceed with caution; (2) measures to prevent intrusion of water moderator into the drain tank, which Crume (1994) reported could make the current configuration critical; and (3) appropriate consideration of the potential value of monitoring neutron multiplication or of adding neutron poison either for normal operation or in response to an upset condition. One potential poison, gadolinium (described further in Appendix D), is a uniquely powerful neutron absorber. The stable compound gadolinium trifluoride (G6F3) would be soluble in, and heavier than, the molten salt and would behave chemically very much like uranium trifluoride (USA. The management of nuclear criticality safety hazards during processing is addressed further in Chapter 8 and Appendix E.