MODELING THE NUCLEAR FUEL CYCLE

MODELING THE NUCLEAR FUEL CYCLE

FUEL CYCLE AND MANAGEMENT

CHRISTOPHER A. JUCHAU,a MARY LOU DUNZIK-GOUGAR,b * and JACOB J. JACOBSON c

KEYWORDS: fuel cycle modeling, discrete material tracking, fuel cycle optimization

a Energy

Solutions, 2345 Stevens Drive, Suite 240, Richland, Washington 99354 Idaho State University and Idaho National Laboratory, 1776 Science Center Drive Suite 335, Idaho Falls, Idaho 83402 c Idaho National Laboratory, 2525 North Fremont Avenue, Idaho Falls, Idaho 83402-3560 b

Received April 10, 2009 Accepted for Publication January 8, 2010

A review of existing analysis codes for nuclear fuel cycle systems was performed to determine if any existing codes meet technical and functional requirements defined for a U.S. national program supporting the global and domestic assessment, development, and deployment of nuclear energy systems. The program would be implemented using an interconnected architecture of different codes ranging from the fuel cycle analysis code, which is the subject of the review, to fundamental physical and mechanistic codes. Four main functions are defined for the code. Function 1 is the ability to characterize and deploy individual fuel cycle facilities and reactors in a simulation while discretely tracking material movements. Function 2 is the capability to perform an uncertainty analysis for each element of the fuel cycle and an aggregate uncertainty analysis. Function 3 is the inclusion of an optimization engine able to optimize simultaneously across multiple objective functions. Function 4

is open and accessible code software and documentation to aid in collaboration between multiple entities and to facilitate software updates. Existing codes, categorized as annualized or discrete fuel tracking codes, were assessed according to the four functions and associated requirements. These codes were developed by various government, education, and industrial entities to fulfill particular needs. In some cases, decisions were made during code development to limit the level of detail included in a code to ease its use or to focus on certain aspects of a fuel cycle to address specific questions. The review revealed that while no two of the codes are identical, they all perform many of the same basic functions. No code was able to perform defined function 2 or several requirements of functions 1 and 3. Based on this review, it was concluded that the functions and requirements will be met only with development of a new code, referred to as GENIUS.

I. INTRODUCTION

of nuclear fuel cycles over a period of time and then to provide output relevant to the parameters of concern. A review of existing nuclear fuel cycle systems analysis codes was performed as part of a Laboratory Directed Research and Development project at the Idaho National Laboratory ~INL!. The goal of the review was to determine if any existing codes met technical and functional requirements defined for a desired nuclear fuel cycle systems analysis tool. As a side note, there are a few existing fuel cycle systems codes that were developed by private entities and are considered proprietary. As such, limited code information was available and review was not possible.

As nuclear power begins to reemerge in the United States and gain momentum in many countries around the world, a number of nuclear fuel cycle systems analysis codes have been developed with the intent to provide policy makers with information about the ramifications of implementing a given nuclear fuel cycle. Parameters of concern include cost, proliferation resistance, waste management, and safety of the proposed nuclear fuel cycle.1 The codes are intended to simulate the evolution *E-mail: [email protected] 136

NUCLEAR TECHNOLOGY

VOL. 171

AUG. 2010

Juchau et al.

Further, one of the defined code requirements was that it be open and accessible. Accordingly, proprietary codes were automatically disqualified. For these reasons, the code review should not be viewed as a comprehensive work. II. DESIRED FUEL CYCLE ANALYSIS TOOL The review of nuclear fuel cycle analysis tools was conducted to determine if an existing code would meet the needs of a new proposed institute called the Simulation Institute for Nuclear Enterprise Modeling and Analysis ~SINEMA!. SINEMA is to be developed as a U.S. national program supporting the global and domestic assessment, development, and deployment of nuclear energy systems. The program would be implemented using an interconnected architecture of different codes ranging from the fuel cycle analysis code that is the subject of the review to fundamental physical and mechanistic codes. The fuel cycle analysis code proposed for SINEMA must contain a transport model that tracks nuclear materials and processes through time. Required outputs of the code include tracking raw materials, processed materials, nuclear material transportation flow, nuclear facilities and their characteristics, material storage, costs and expenditures, proliferation implications, and other products and by-products such as electricity and hydrogen production. The source and sink terms for the code will be supplied by other codes in SINEMA that model specific portions of the fuel cycle in more detail. III. TECHNICAL AND FUNCTIONAL REQUIREMENTS The defined technical and functional requirements are a product of multiparty ~national laboratories, academia, and industry! collaboration. Code requirements are organized as subsets of four main functions. Function 1 is the ability to characterize and deploy individual fuel cycle facilities and reactors in a simulation. The first function also calls for the code to track material movements in a simulation in a discrete manner, for example, at the fuel batch or fuel assembly level. The discrete tracking of materials is required for simulation of parameters such as ~a! the selection of used fuel assemblies for processing into a new fuel of appropriate isotopic composition ~e.g., for mixed-oxide or fast reactor fuel! and ~b! dose and heat load generated by specific assemblies during shipping and storage. Detailed isotopic tracking is required to address advanced fuel cycle viability questions. For example, isotopes of plutonium and higher actinides in spent nuclear fuel ~SNF! will be present in different quantities, both relative and absolute, depending on the different batches of SNF. It will be necessary to mix batches of the recycle materials to create the required mix of fissile, fertile, and other isotopes for new fuel assemblies. The NUCLEAR TECHNOLOGY

VOL. 171

AUG. 2010

MODELING NUCLEAR FUEL CYCLES

viability of recycling higher actinides depends upon developing practical methods to mix very different batches of recycle fissile material. A detailed understanding and accounting of SNF isotopics is required to conquer the significant challenge of creating uniform feeds, especially of fissile material, for fuel fabrication. Also inherent in Function 1 is the ability to account for radioactive decay as part of material tracking. Doing so is necessary for evaluating the above listed parameters. The identification, tracking, and where applicable, calculation of radioactive decay of 81 isotopes and chemical elements should be sufficient for most analyses.2 Tracking and accounting for the decay of the roughly 2200 radionuclides that could be present in a simulation would be very computationally expensive. Function 2 is the capability to perform an uncertainty analysis for each element of the fuel cycle and an aggregate uncertainty analysis. Function 3 requires the inclusion of an optimization engine able to optimize simultaneously across multiple objective functions. Function 4 states that code software and documentation must remain open and accessible to aid in collaboration between multiple entities and facilitate software updates. Table I summarizes the requirements for each technical function. IV. CODE REVIEW The result of the code review with respect to the defined technical functions and requirements was a conclusion that creation of a new code was necessary. A prototype of the new code, titled GENIUS 1, was constructed by the authors and is included in the review. Table II gives a summary of the code comparison results. Boxes marked with a check indicate that the code in question fulfills a particular requirement. The codes reviewed for this paper share a few basic functions. All of the codes simulate the deployment of reactors over time to meet a defined energy demand curve and include the capability to simulate the transition from an existing fuel cycle to a proposed future nuclear fuel cycle. The codes quantify the necessary mass of material needed to fill reactor fuel orders at each step of the fuel cycle throughout a simulation. In the following sections, codes are organized according to one key feature, and distinctive features are described. Conceptually, each code belongs to one of two groups based on how material movement is simulated: annualized fuel tracking codes and discrete fuel tracking codes. V. ANNUALIZED FUEL TRACKING CODES Annualized fuel tracking codes track material movement in a simulation as annual average mass flows. The reviewed codes in this category are Code for Advanced 137

Juchau et al.


TABLE I Requirements Organized According to Four Technical Functions Function or Requirement

Description

Function 1 R1.1 R1.2 R1.3 R1.4 R1.5 R1.6 R1.7 R1.8 R1.9

Characterize and deploy individual fuel cycle facilities and reactors Simulations must be able to reflect all significant design data for elements of a fuel cycle. Track quantities of natural resources as a function of time, location, and accessibility. Track process materials as a function of time, location, and accessibility ~element and isotope!. Track the operations status of each production facility as a function of time, location, and capacity. Track the operations status of each storage facility as a function of time, location, and capacity. Track the operations status of each disposal facility as a function of time, location, and capacity. Track status of nuclear materials transportation as a function of time, location, and type. Track products and by-products as a function of time, location, and type ~electricity, heat, hydrogen, etc.!. Track costs0expenditures as a function of time, location, and type.

Function 2 R2.1 R2.2 R2.3

Perform component and aggregate uncertainty analyses Capable of performing uncertainty analysis for each element of the fuel cycle. Generate sensitivity coefficients for each element of the fuel cycle. Capable of performing aggregate uncertainty analysis by propagating the uncertainty in each element of the fuel cycle.

Function 3 R3.1

Optimize simultaneously across multiple objective functions Capable of running in a semiautomated mode using inputs to produce desired model outputs. ~e.g., nonproliferation, economic, and waste management targets!. Be able to dynamically perturb local optimum solutions to test robustness and adaptability. Support a graphical user interface.

R3.2 R3.3 Function 4 R4.1 R4.2 R4.3

Open and accessible code software and documentation Maintain abstractions between data and process algorithms. Open and accessible code architecture, source code, and documentation. Be able to communicate with other codes through weak links0databases.

TABLE II Summary of Code Comparison Results Requirement

CAFCA

R1.1 R1.2 R1.3 R1.4 R1.5 R1.6 R1.7 R1.8 R1.9

COSI

CEPMNFC

DANESS

NFCSim

⻬⻬⻬⻬⻬

⻬⻬

⻬⻬⻬⻬⻬

⻬⻬

⻬⻬

⻬⻬

VEGAS

VISION

NFCSS

GENIUS 1

⻬⻬⻬

R2.1 R2.2 R2.3 R3.1 R3.2 R3.3 R4.1 R4.2 R4.3

138

⻬⻬⻬

⻬

⻬⻬

⻬

⻬⻬

⻬

⻬

⻬⻬

⻬

NUCLEAR TECHNOLOGY

VOL. 171

⻬⻬

AUG. 2010

Juchau et al.

Fuel Cycles Assessments ~CAFCA!, Dynamic Analysis of Nuclear Energy Systems Strategies ~DANESS!, VEGAS ~not an acronym!, Verifiable Fuel Cycle Simulation ~VISION!, and Nuclear Fuel Cycle Simulation System ~NFCSS!. It is important to note that while several of the codes in this category are capable of performing a wide variety of analyses, they do not meet most of the GENIUS Function 1 requirements because they do not discretely track materials in a simulation. The CAFCA code has been developed by successive generations of graduate students at the Massachusetts Institute of Technology. CAFCA is implemented in Matlab. The goal of a CAFCA simulation is to minimize the transuranic ~TRU! inventory of a given scenario. Uranium and TRUs are tracked as a group. Fission products are not tracked.3 The current version of CAFCA can be used to simulate multiple different fuel cycles and three separate geographical regions with different power demands.4 Deployment of separations and reprocessing facilities in a simulation are optimized so that a minimum capacity factor is maintained for the lifetime of each plant. Separations facilities remove uranium from used nuclear fuel; reprocessing facilities remove TRUs from used nuclear fuel. CAFCA performs economic analyses for a simulation calculating the average total cost of electricity for a scenario. Radioactive decay is not calculated but may be included in future versions of the code.5 Front end fuel cycle capacity is assumed to be a nonconstraining factor in CAFCA and is not modeled in the code.4 DANESS is under development at Argonne National Laboratory ~ANL!. Ten different reactor-fuel combinations can be simulated in a scenario. Reactors and facilities are characterized by type, not individually. New and used fuel isotopic recipes are entered by the user via the graphical user interface. Material flows are tracked isotopically in a simulation, and radioactive decay is incorporated. DANESS calculates a levelized cost of electricity ~e.g., cost per kilowatt hour! for a scenario. A waste fee may be included in cost calculations. This is a fee similar to the mil0kilowatt hour fee charged to utilities to fund the development of a deep geologic repository for used nuclear fuel. A user may rely on an economic decisionmaking model included to make deployment decisions.6 The VEGAS code was developed at Los Alamos National Laboratory ~LANL! for the purpose of performing simplified simulations in order to identify promising scenarios for more detailed analysis using NFCSim ~a discrete flow code!. The primary goal of a VEGAS simulation is to minimize TRU material. The deployment of reprocessing facilities is optimized to maintain a minimum capacity factor for each plant. Reactors and fuel cycle facilities are characterized by type in a VEGAS simulation. Radioactive decay is not incorporated. An annual levelized cost of electricity is calculated for a simulation.7 VISION is the U.S. Department of Energy’s Advanced Fuel Cycle Initiative nuclear fuel cycle systems code. The code was initially developed at ANL as the NUCLEAR TECHNOLOGY

VOL. 171

AUG. 2010


DYMOND code and is currently undergoing development at INL with collaboration from ANL, Sandia National Laboratories, Idaho State University, North Carolina State University, and the University of Wisconsin. VISION has been constructed using the PowerSim software package. The code is capable of modeling nuclear fuel cycle scenarios using multiple reactor and fuel combinations. New and used fuel recipes are calculated exogenously and included with the code package. The code calculates radioactive decay and tracks material flows isotopically.8 Individual reactors and fuel cycle facilities may be deployed in a simulation. An economics package called VISION.ECON is used to calculate the levelized cost of electricity for a simulation.9 NFCSS was developed at the International Atomic Energy Agency ~IAEA!. Reactors and fuel cycle facilities are characterized by type in NFCSS. Inputs included with the code are a mix of historical data from IAEA databases and reports from various entities such as the IAEA PRIS reactor database and other publications and consultant reports. Materials are tracked by isotope, and radioactive decay is incorporated in a simulation.10 New and used fuel isotopic recipes are calculated prior to simulation by a burnup and depletion engine. Four different reprocessing scenarios can currently be simulated. Access to a full version of the NFCSS is available online to member states following successful processing of a request form. Economic and environmental analyses packages may be added in the future.11

VI. DISCRETE FUEL TRACKING CODES The second group of codes tracks material movement in discrete fuel batches or fuel assemblies. Codes reviewed in this category are Comellini Sicard ~COSI!, Comprehensive Economic and Physical Model of the Nuclear Fuel Cycle ~CEPMNFC!, Global Evaluation of Nuclear Infrastructure and Utilization Scenarios ~GENIUS!, and Nuclear Fuel Cycle Simulator ~NFCSim!. COSI is under development by Commissariat à l’Énergie Atomique in France. The code can simulate multiregional scenarios incorporating individually defined reactors and fuel cycle facilities. The code can also simulate the operation of multinational facilities. COSI is composed of several interconnected codes, including a cycle-by-cycle burnup and depletion engine.12 The code can be used to provide an economic analysis of a nuclear fuel cycle including calculations for levelized cost of electricity. COSI can also be used to quantify the number of waste packages required for a given simulation and analyze the impact a given waste form may have on the integrity of those containers.13 The CEPMNFC code was developed as part of a PhD dissertation at Cornell University.14 Individually characterized reactors may be included in a simulation. The 139

Juchau et al.


CEPMNFC code also incorporates a cycle-by-cycle burnup and depletion engine to calculate new and used fuel isotopic compositions. Actinides are tracked at the isotopic level, while fission products are tracked as a group. Radioactive decay is incorporated. The code calculates the levelized cost of electricity for a given scenario.14 The GENIUS 1 code was developed at INL. GENIUS 1 is a multiregional simulator in which individually characterized reactors and fuel cycle facilities may be included in a simulation. Materials are tracked isotopically. New and used fuel recipes have been calculated exogenously and are included in the code. GENIUS 1 can be used to simulate three basic fuel cycle arrangements: a fuel leasing agreement between two or more countries, a market-based system, or a nationalized system in which a given region obtains all fuel cycle services in-house.15 Further development ~GENIUSv2! is currently taking place at the University of Wisconsin. In GENIUSv2, facilities can be further categorized according to the institutions that own them, such as a utility or government entity, and rules can then be assigned that govern the trade of materials between those institutions. Facilities in GENIUSv2 can also be assigned financial parameters such as tax and interest rates for the purpose of economic analysis. Front-end fuel cycle facilities will have the option of requesting recycled material from separations plants, and radioactive decay of uranium, TRUs, and fission products of interest will be incorporated in GENIUSv2 ~Ref. 16!. NFCSim was developed at LANL as part of the U.S. Department of Energy’s Advanced Fuel Cycle Initiative. The code was developed primarily to explore different fuel cycle options for the United States. Individually characterized reactors, fuel cycle facilities, and acceleratordriven systems can be deployed in a simulation. Material flows are tracked isotopically and radioactive decay is calculated. A proposed schedule for shipping used fuel to a repository can be included in an NFCSim simulation. The code calculates a levelized cost of electricity for a modeled scenario.7

VII. CONCLUSION While no two of the reviewed codes are identical, they all perform many of the same basic functions. For example, all of the codes deploy reactors to meet a nuclear energy demand curve. After reactor deployment, the codes attempt to quantify the response of a given nuclear fuel cycle to the fuel ordered by reactors in a simulation. All of the codes simulate, to some extent, both open and closed fuel cycles. The existence of codes with overlapping functions is largely a result of reinventing the proverbial wheel. Because many of the codes are at least partly proprietary, 140

persons or organizations having need of such modeling capabilities have resorted to developing their own. Ideally, fewer restrictions on use, combined with a more collaborative attitude toward development, would have resulted in fewer codes with more capabilities. Several of the proprietary codes have lost funding and are no longer used. Unfortunately, in some cases, implemented restrictions do not allow continued development by another organization. There is ongoing debate regarding the usefulness of modeling the nuclear fuel cycle in a discrete manner. It is possible that uncertainties inherent in the modeling process will eclipse any gain in fidelity from discrete tracking. Conversely, it can be argued that approximating fuel cycle material flows as annualized mass flows is an oversimplification resulting in output too general for analyses involving parameters such as transportation. The debate over the utility of modeling the nuclear fuel cycle in a discrete manner is one part of the larger question, “To what level of detail is modeling realistic?” For example, Table I lists for Function 1, Requirement 1.1 the desire “to reflect all available design data for elements of a fuel cycle.” Reflecting all available data elements of a fuel cycle in a simulation is likely not possible. Developers of nuclear fuel cycle simulation tools will always find it necessary to choose a level of detail that balances model credibility with usefulness.

REFERENCES 1. “Report to Congress—Advanced Fuel Cycle Initiative: Objectives, Approach, and Technology Summary,” U.S. Department of Energy, Office of Nuclear Energy, Science, and Technology ~May 2005!. 2. S. PIET, “Selection of Isotopes and Elements for Fuel Cycle Analysis,” Proc. Int. Topl. Mtg. Advances in Nuclear Fuel Management IV (ANFMIV), Hilton Head Island, South Carolina, April 12–15, 2009, American Nuclear Society ~2009!. 3. T. BOSCHER, P. HEJZLAR, M. S. KAZIMI, N. E. TODREAS, and A. ROMANO, “The CAFCA Code for Simulation of Nuclear Fuel Cycle: Description of Methodology, Assumptions, and Initial Results,” MIT-NFC-TR-069, Center for Advanced Nuclear Energy Systems, Massachusetts Institute of Technology ~2004!. 4. J. WARBURTON, “Determination of the Proper Operating Range for the CAFCA IIB Fuel Cycle Model,” MS Thesis, Massachusetts Institute of Technology ~2007!. 5. T. BOSCHER, P. HEJZLAR, M. S. KAZIMI, N. E. TODREAS, and A. ROMANO, “Alternative Fuel Cycle Strategies for Nuclear Power Generation in the 21st Century, Revision 1,” MIT-NFC-TR-070-REV1, Center for Advanced Nuclear Energy Systems, Massachusetts Institute of Technology ~2005!. 6. L. VAN DEN DERPEL, A. YACOUT, D. WADE, and H. KHALIL, “DANESS—Dynamic Analysis of Nuclear Energy NUCLEAR TECHNOLOGY

VOL. 171

AUG. 2010

Juchau et al.

Systems Strategies,” Argonne National Laboratory ~June 2005!; http:00www.daness.anl.gov0documents0. 7. C. G. BATHKE, E. A. SCHNEIDER, S. F. DeMUTH, and M. R. JAMES, “Report of LANL Advanced Fuel Cycle Systems Analyses for FY 2003,” LA-UR-03-8740, Los Alamos National Laboratory ~2003!. 8. A. M. YACOUT et al., “Vision—Verifiable Fuel Cycle Simulation of Nuclear Fuel Cycle Dynamics,” presented at Nuclear Waste Management Sypmosium ~WMS!, Tucson, Arizona, February 26–March 2, 2006. 9. J. JACOBSON, Idaho National Laboratory, Personal Communication ~Aug. 2007!. 10. “Integrated Nuclear Fuel Cycle Information System ~iNFCIS!,” International Atomic Energy Agency, http:00wwwnfcis.iaea.org0 11. M. CEYHAN, “Nuclear Fuel Cycle Simulation System ~VISTA!,” IAEA-TECDOC-1535, International Atomic Energy Agency, Vienna ~2007!. 12. J. GROUILLER, G. GLAMENBAUM, B. SICARD, M. MUS, J. MARTIN, J. DEVEZEAUX DE LAVERGNE, and O. COMELLINI, “COSI, A Simulation Software for a Pool of

NUCLEAR TECHNOLOGY

VOL. 171

AUG. 2010


Reactors and Fuel Cycle Plants: Applications to the Study of the Deployment of Fast Breeder,” presented at Int. Conf. Fast Reactors and Related Fuel Cycles, Atomic Energy Society of Japan ~1991!. 13. L. BOUCHER and J. GROUILLER, “COSI: The Complete Renewal of the Simulation Software for the Fuel Cycle Analysis,” Proc. 14th Int. Conf. Nuclear Engineering, Miami, Florida, July 17–20, 2006, v.1, p. 865, American Society of Mechanical Engineers ~2006!. 14. E. A. SCHNEIDER, “A Comprehensive Physical and Economic Model of the Nuclear Fuel Cycle,” PhD Dissertation, Cornell University ~2002!. 15. C. A. JUCHAU, “Development of the Global Evaluation of Nuclear Infrastructure Utilization Scenarios Nuclear Fuel Cycle Simulation Tool,” MS Thesis, Idaho State University ~2008!. 16. K. M. OLIVER, P. P. H. WILSON, A. REVEILLERE, T. W. AHN, K. DUNN, K. HUFF, and R. ELMORE, “Studying International Fuel Cycle Robustness with the GENIUSv2 Discrete Facilities0Materials Fuel Cycle Analysis Tool,” presented at Int. Conf. GLOBAL 2009, Paris, France, September 6–11, 2009.

141