Engineering topics
Publications and source records attributed to Stover, Tracy.
Overview of NNSA Projects (Criticality Safety) – Savannah River Plutonium Processing Facility – Surplus Plutonium Disposition Project – Tritium Finishing Facility – Progress of Criticality Safety Analysis – Processes requiring CAAS
The validation and bias quantification methods employed by the Criticality Safety Engineering group at Savannah River Nuclear Solutions are discussed. Both SCALE and MCNP are validated against benchmark experiments, where these benchmark models have been divided into subgroups representative of the types of fissile material operations at SRNS. A quantification of the bias is made for each subgroup for each code and the three methods employed for quantification are discussed. The bias information is employed directly in the establishment of the Upper Safety Limit (USL) value of the neutron multiplication factor when using standardized documented SRNS methods. A simple example of validation and bias quantification is given and the procedure for setting the USL using the derived bias information is also presented.
As one of its missions, the Savannah River Site seeks to be a leader in the safe and productive disposition of special nuclear materials. One of the premier programs is the disposition of used nuclear fuel, primarily from research reactors, through dissolution, purification, and down blend into low enriched uranium for commercial power reactors. Prior to use, this fuel contained highly enriched uranium (HEU). In an effort to advance this expertise, this study has been performed to demonstrate the feasibility to disposition a number of special geometry HEU fuels. These fuels come from a number of different domestic and foreign research reactors which had geometries considerably different from round or square bundles of plates or rods. As such, they cannot be dispositioned in current H-Canyon dissolver inserts so a new, multi-function dissolver insert may be considered. While this is not a current mission, it is believed that this study's scoping calculations demonstrate the feasibility of the site taking such a mission. At the H-Canyon facility, a dissolver is a large stainless steel pot, loaded with a nitric acid solution into which uranium-alloy fuel is dissolved. The dissolver uses a device called an insert to feed the fuels to the acid solution in a controlled manner. Most fuel forms fit into one of the existing inserts which include one with ten {approx}5-inch diameter circular wells, one with four {approx}4.5 inch by {approx}15 inch slab wells, and one specifically designed for the Oak Ridge National Laboratory High Flux Isotope Reactor (HFIR). This study considered five bounding fuel candidates for scoping the potential new insert. The fuels include a generic annular plate fuel (APF), the Tower Shield Reactor (TSR) fuel elements, the Reactor A Haute Flux (RHF) core, Knolls Atomic Power Laboratory (KAPL) reactor fuel, and SLOWPOKE cores. These are expected to bound a range of proposed fuel candidates. Scoping studies were made to establish geometric restrictions for the new dissolver insert and to determine if these fuels could safely be dissolved. Homogenous models were used for the scoping study due to time constraints. Computational modeling was performed using KENOVI in the SCALE 6.1 code package.
An approach new to the H-Canyon facility criticality safety analysis has been performed using highly detailed models of the H-Modified (HM) process mixer settlers. The HM process uses tributyl phosphate (TBP) in organic n-paraffin to extract and purify uranium from used nuclear fuel. This occurs in two cycles of counter-current solvent extraction. The aqueous stream is a weak to moderate nitric acid and the feed stock is an aqueous stream. The first Cycle, consisting of two 16-stage banks (A and B-Banks) and one 12-stage bank (C-Bank), is a three step process that separates uranium from the fuel matrix, fission products, and most of the transuranic actinide content. The first Cycle aqueous product is concentrated, adjusted for acidity and fed to the second Cycle. The second Cycle, consisting of one 16-stage bank (D-Bank) and one 12-stage bank (E-Bank), is a two-step process that further purifies the uranium and removes to a great extent remaining actinides and fission products Each stage of the mixer settler banks is modeled as three unique compositions: the mixer as a combined composition of aqueous and organic components and the settler one aqueous and one organic component. This work addresses what is believed to be normal, credible abnormal, and beyond credible abnormal conditions for a particular postulated series of process parameter upsets. The level of detail and the source for the stage-wise compositions are both new and unique to criticality safety analyses of the HM process. Initial work was presented in Ref. 1 and has since been refined. A predictive chemistry model for flowsheet development has been used to solve the solvent extraction equations for the prescribed conditions and the output of that model was then transformed into compositions for each mixture in each stage. The remaining question upon which to achieve consensus is if the approach was appropriately conservative by concluding the chemistry is appropriately bounded. Computational modeling for criticality calculations was performed using KENO-VI in the SCALE 6.1 code package. Computational modeling of the chemistry was performed using the SEPHIS Modification 4 code.
Savannah River Nuclear Solutions has tasked its H Canyon facility to prepare for disposition of used Oak Ridge National Laboratory High Flux Isotope Reactor (HFIR) cores. HFIR cores consist of an inner and outer assembly of highly enriched uranium fuel plates. The two elements are annular, nest inside each other, and the fuel is a curved plate held between two collars. The H-Canyon facility has a dissolver insert specifically designed for HFIR cores and has performed this type of campaign before between 1976 and 1988. At H-Canyon, a dissolver is a large stainless steel pot loaded with a nitric acid solution into which uranium-alloy fuel is dissolved. The dissolver uses a device called an insert to feed the fuels to the acid solution in a controlled manner. The mission will be to dissolve a number of HFIR cores, purify the uranium, and down blend it to low enriched uranium suitable for commercial power reactors. HFIR has a sister core. A French high flux research reactor, called Reactor A Haut Flux (RHF), has nearly the same geometry as the outer element of HFIR. The annulus is only slightly smaller and the RHF active fuel height is {approx}97 cm while the HFIR active fuel height is only {approx}51 cm. SRNS is currently storing RHF cores (repatriated through circa 2003) awaiting a path for disposition. There is now an opportunity. RHF cores fit in the HFIR dissolver insert. Dispositioning them concurrent with HIFR avoids a future facility evolution of just to change dissolver inserts again. This work is the result of an initiative to propose appending the RHF cores to the anticipated HFIR campaign. Geometry and chemistry differences are evaluated and criticality safety assessed. The RHF cores were semi-explicitly modeled in various stages of dissolution. Normal and credible abnormal conditions were assessed. RHF was shown to be bounded by the dissolution of a complete HFIR core in both neutron multiplication and dissolver chemistry. Computational modeling was performed using KENOVI in the SCALE 6.1 code package.
A reactor core design engineer, even one with many years of experience or one using advanced fuel pattern optimization software, will likely fine tune the pattern before finalizing the design. Fine tuning can further minimize local enthalpy peaks, optimize reactivity to extend burnup, and add additional margin to limits on reactivity coefficients or heat flux. For reactors that have reached an equilibrium fuel cycle, the number of feeds each cycle is typically the similar to previous cycles as is their placement in the core. It is then the engineer's responsibly to efficiently select the placement of burned assemblies to ensure optimal fuel utilization, satisfy safety limits, and meet cycle energy requirements. One of the engineer's most commonly used tools in this fine tuning investigation is to swap the positions of two assemblies, most likely two previously irradiated ('burned') assemblies since their reactivity is lower than fresh fuel. Such an operation is referred to in this work as a binary perturbation or a 'swap'. If quarter core symmetry is maintained, the perturbation translates to the other three quadrants. Likewise if eighth core symmetry is maintained, the perturbation translates to other seven octants. The question is then, how much of the core design space can binary perturbations possibly span? This work seeks to quantitatively define the space spanned by m number of binary perturbations as a function of variable size through combinatorial principles and computer simulation. The results of this study are useful in understanding the in-core fuel management problem faced by both professionals who employ hand manipulations and those who attempt stochastic solutions of the problem.
Savannah River Nuclear Solutions has tasked its H-Canyon facility to prepare for disposition of used Oak Ridge National Laboratory High Flux Isotope Reactor (HFIR) cores. HFIR cores consist of an inner and outer assembly of highly enriched uranium fuel plates. The two elements are annular, nest inside each other, and the fuel form is a curved plate held between two collars. H-Canyon has a dissolver insert specifically designed for HFIR cores and has performed this type of campaign before between 1976 and 1988. At the H-Canyon facility, a dissolver is a large stainless steel pot loaded with a nitric acid solution into which uranium-alloy fuel is dissolved. The dissolver uses a device called an insert to feed the fuels to the acid solution in a controlled manner. The mission will be to dissolve a number of HFIR cores, purify the uranium, and down blend it to low enriched uranium suitable for commercial power reactors.
An approach new to the H-Canyon facility criticality safety analysis has been performed using highly detailed models of the H-Modified process mixer settlers. The HM process uses tributyl phosphate (TBP) in organic n-paraffin to extract and purify uranium from used nuclear fuel. This occurs in two cycles of counter-current solvent extraction. The aqueous stream is a weak to moderate nitric acid and the feed stock is an aqueous stream. The 1st Cycle, consisting of two 16-stage banks (A and B-Banks) and one 12-stage bank (C-Bank), is a three step process that separates uranium from the fuel matrix, fission products, and most of the transuranic actinide content. The 1 st Cycle aqueous product is concentrated, adjusted for acidity and fed to the 2 nd Cycle. The 2 nd Cycle, consisting of one 16-stage bank (D-Bank) and one 12-stage bank (E-Bank), is a two step process that further purifies the uranium and removes to a great extent remaining actinides and fission products.
The criticality safety of Savannah River Site's H Canyon two cycle uranium extraction and purification process, called the HM process, is examined. Explicit Monte Carlo models of the multi-stage mixer-settlers are generated for the first time using the KENO-VI module of the SCALE 6.1 code package and spanning normal operating conditions and a wide range of potential upset conditions with stagewise compositions based directly on SEPHIS Mod-4 chemistry models of the same conditions. To achieve industrial scale processing, mixer-settlers are often not geometrically favorable to prevent criticality so other means must be used for criticality safety, namely control of fissile concentration and mass. Historically, the worst case upset is a reflux condition resulting from a stream flow rate, temperature, or composition that is off from the normal operating conditions. A reflux is considered to be the state in which the fissile material is not proceeding in the desired direction, i.e. is stripping rather than extracting or vice versa. The criticality safety analyses which set the current operational limits were conducted many years ago with less sophisticated codes and only loosely tied criticality to chemistry. Those analyses conservatively assumed that a reflux changed the behavior of the mixer-settler such that the fissile content collected continuously in only three stages of the mixer-settler, at a rate based on the SEPHIS modeling, until criticality occurred. This implies that under some condition the distribution coefficient of the fissile material either goes to infinity or zero, which is not physically achievable. In such a scenario, the operator would have some amount of time to react and correct the problem before sufficient fissile material built up in the mixer-settler and caused a criticality. In reality, an upset in a stream would cause the process to transition to a new, though undesired, equilibrium state of operation. The distribution coefficients would change, but would not diverge. In this work, aqueous and organic streams are modeled for both the mixer and the settler region of each stage of each bank using chemical distribution data from SEPHIS and solving the chemical equations for a wide range of operational parameters spanning normal and upset conditions. Conservative assumptions are made regarding enrichment and reflection and the whole bank multiplication factor is evaluated using discrete stage-wise compositions. (authors)
At the H-Canyon outside facilities, highly enriched uranyl nitrate solutions collected from leaks or during deliberate maintenance activities is collected into 5-gallon polyethylene bottles. The current criticality safety evaluation for the facility postulates that a criticality event is possible if the solutions are allowed to concentrate through evaporation, freezing, storing multiple bottles in a geometrically unfavorable array, or any combination of these conditions. Using chemical data on uranyl nitrate freezing from Savannah River National Laboratory, protected limits on concentration and enrichment, and normal operational restrictions, criticality was shown to be incredible under all normal and credible abnormal conditions. A bounding normal solution is analyzed under normal conditions and shown to be substantially subcritical. A bounding credible abnormal solution is defined as 9.2 g U{sup 235}/L and 73 wt. % enriched and 0 M excess nitric acid. The concentration and enrichment are limitations already protected by the facility for other processes. Because acidity is not specifically restricted, the most reactive lower bound is selected. The credible abnormal worst placement is in a concrete basin reflected on three sides. There the bottle(s) would be subject in the worst cases to flooding (full water reflection) and postulated concentration through evaporation or freezing. Computational modeling was performed using KENOVI in the SCALE 6.1 code package. (authors)
This work represents the initial phase of an overall effort to reduce the analytical burdens and increase flexibility of the K-Area Complex which provides stewardship of special nuclear materials. A hand calculation method is being extended to apply to large arrays of highly heterogeneous shipping packages. This method has the potential to quickly determine the impact of perturbations on the array. The objective of the particular work reported herein is twofold. First, the analytical results upon which the Limiting Surface Density (LSD) method was originally based were reproduced using the modern computational code KENO-VI in SCALE 6.1. The second objective was to test the classic LSD method on storage of large arrays of highly heterogeneous shipping packages. Specifically, the objective is to determine if the method is adequate for existing arrays of 9975 shipping packages with fissile metal content. The classic LSD method was derived for simple packages of bare spheres of fissile material centered in an air filled cubic unit cell and placed in a cubic-pitch water-reflected array, the properties of which could then be translated to more practical array and package geometries. The classic method did not address highly heterogeneous packages which are close to neutronically decoupled from other identical packages. Therefore, the material based array constant (called c{sub 2}) as defined by the classic method is determined specifically for the packages in question. (authors)
At the H-Canyon outside facilities, highly enriched uranyl nitrate solutions collected from leaks or during deliberate maintenance activities is collected into 5-gallon polyethylene bottles. The current criticality safety evaluation for the facility postulates that a criticality event is possible if the solutions are allowed to concentrate through evaporation, freezing, storing multiple bottles in a geometrically unfavorable array, or any combination of these conditions. Using chemical data on uranyl nitrate freezing from Savannah River National Laboratory, protected limits on concentration and enrichment, and normal operational restrictions, criticality was shown to be incredible under all normal and credible abnormal conditions.
The criticality safety of Savannah River Site’s H Canyon two cycle uranium extraction and purification process, called the HM process, is examined. Explicit Monte Carlo models of the multi-stage mixer-settlers are generated for the first time using the KENO-VI module of the SCALE 6.1 code package and spanning normal operating conditions and a wide range of potential upset conditions with stage wise compositions based directly on SEPHIS Mod-4 chemistry models of the same conditions.
This work represents the initial phase of an overall effort to reduce the analytical burdens and increase flexibility of the K-Area Complex which provides stewardship of special nuclear materials. A hand calculation method is being extended to apply to large arrays of highly heterogeneous shipping packages. This method has the potential to quickly determine the impact of perturbations on the array.
As one of its missions, the Savannah River Site seeks to be a leader in the safe and productive disposition of special nuclear materials. One of the premier programs is the disposition of used nuclear fuel (UNF), primarily from research reactors, through dissolution, purification, and down blend into low enriched uranium for commercial power reactors. Prior to use, this fuel contained highly enriched uranium (HEU). In an effort to advance this expertise, this study has been performed to demonstrate the feasibility to disposition a number of special geometry HEU fuels. These fuels come from a number of different domestic and foreign research reactors which had geometries considerably different from round or square bundles of plates or rods. As such, they cannot be dispositioned in current H-Canyon dissolver inserts so a new, multi-function dissolver insert may be considered. While this is not a current mission, it is believed that this study’s scoping calculations demonstrate the feasibility of the site taking such a mission.