PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future The Westin Miyako, Kyoto, Japan, September 28 – October 3, 2014, on CD-ROM (2014)
DEVELOPMENT OF COMPUTER CODE PACKAGES FOR MOLTEN SALT REACTOR CORE ANALYSIS Y. Jeong, S. Choi, and D. Lee* School of Mechanical and Nuclear Engineering Ulsan National Institute of Science and Technology, Ulsan, Republic of Korea
[email protected] [email protected] [email protected]* ABSTRACT This paper presents the implementations of the Oak Ridge National Laboratory (ORNL) approach for Molten Salt Reactor (MSR) core analysis [1] with two nuclear reactor core analysis computer code systems. The first code system has been set up with the MCNP6 Monte Carlo code [2], its depletion module CINDER90 and the PYTHON script language. The second code system has been set up with the NEWT transport calculation module and ORIGEN depletion module connected by TRITON sequence in SCALE code [3], and the PYTHON script language. The PYTHON script language is used for implementing the online reprocessing of molten-salt fuel, and feeding new fertile material in the computer code simulations. In this paper, simplified nuclear reactor core models of a Molten Salt Breeder Reactor (MSBR), designed by ORNL in the 1960’s [4], and FUJI-U3 [5] designed by Toyohashi University of Technology (TUT) in the 2000’s, were analyzed by the two code systems. Using these, various reactor design parameters of the MSRs were compared, such as the multiplication factor, breeding ratio, amount of material, total feeding, neutron flux distribution, and temperature coefficient. Key Words: MSR, molten-salt, online processing, MCNP, SCALE
1. INTRODUCTION Thermal spectrum molten salt reactors (MSRs) consist of fuel-bearing molten-salt fluid (e.g, LiF - BeF2 - ThF4 - UF4) which acts as a coolant as well, and the bare graphite moderator. The fluid fuel is a unique characteristic of MSRs, distinguished from other nuclear reactor concepts by giving special properties. The molten-salt fuel makes it possible to do online fuel reprocessing & refueling, meaning MSRs can operate long term without stopping and can achieve excellent neutron economy. This fuel does not need to be fabricated, so that the MSRs are also beneficial in terms of economics. There is also a high level of inherent safety, due to its large negative temperature coefficient stem from the large temperature expansion coefficient. In terms of the nuclear fuel cycle, the thorium cycle produces very small amounts of plutonium and minor actinides (MAs) compared to the conventional uranium cycle, so it can increase proliferation resistance, while making the MSR operate as a breeder reactor. MSRs can use spent fuel as a converter reactor, which means it uses radioactive waste from Pressurized Water Reactor (PWR) or others. Therefore MSRs can operate for a variety of purposes.
*
Reactorcore.unist.ac.kr
Y. Jeong, S. Choi & D. Lee
Because of those special concepts, the MSR was selected as one of the gen-IV (generation-IV) design by United States Department of Energy (DOE). However the development of the MSR has almost been stopped because of relatively little related technology, less operating experience, and a lack of computational code for analyzing MSR cores compared to other gen-IV designs. For the development of MSR research, this paper provides the validity and suitability of existing computer code systems for this reactor. One is the MCNP6, a probabilistic Monte Carlo computer code, which has CINDER90 for depletion calculation, using ENDF-VII nuclear data. The other is SCALE, a deterministic computer code, which has transport solver NEWT and the depletion module ORIGEN, also using ENDF-VII nuclear data. The TRITON sequence obtains the reaction rate from NEWT and gives it to ORIGEN for the depletion calculation. Both computer codes have been used for analyzing many kinds of nuclear reactor core such as PWR, CANDU. Contemporary computer codes like MCNP6 or SCALE are good for fixed solid fuel core systems. However, due to the molten-salt fuel, MSR analysis needs more functions, such as online reprocessing & refueling, and circulating fuel. J.J. Power of ORNL suggested in 2013 a method for simulating the MSBR with SCALE, which does not support continuous material processing. In order to simulate the characteristics of MSR, the method proposes dividing a depletion time into short time intervals and batchwise reprocessing & refueling at each step, by using SCALE code. This paper follows J.J Power’s method and applies to MSBR by using MCNP6 and SCALE code. Furthermore FUJI-U3 core is analyzed by using MCNP6 code. However the other characteristic, circulating molten-salt fuel, which causes delayed neutron precursor drift is not considered in this paper. Section 2 explains how the new approach is implemented in the two code systems; MCNP6 code & PYTHON script system, and NEWT-TRITON-ORIGEN codes & PYTHON script system. Each MSR model has distinct reprocessing capabilities and different objectives. Section 3 shows which unit cell model is used to verify the two code system. Specific information of each unit cell is stated in Section 3 such as geometry, material composition, power level, etc. Section 4 analyzes the results of depletion calculation of a two unit cell by the two nuclear core analysis computer code systems. It contains various parameters to analyze the MSR core model such as the multiplication factor, breeding ratio, change of amount of fuel, amount of fuel feeding, temperature coefficient, and neutron flux distribution. 2. REPROCESSING METHOD Contemporary computer codes like MCNP6 or SCALE do not support online reprocessing & refueling. In order to simulate MSR characteristics, realization of the function is necessary. The function can be realized by dividing a depletion time into short time interval and batchwise reprocessing & refueling at each step which makes the similar with real online reprocessing & refueling. Overall procedures are repeated to take material information after the depletion, during the short time interval, and make a new input by using PYTHON script language. The material processing is carried out by PYTHON in between each interval. Each MSR reactor design has several different criteria because of its requirements and capacities. From next chapter, the reprocessing & refueling methods of implementing two MSR core designs are introduced.
2 / 11
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
DEVELOPMENT OF COMPUTER CODE SYSTEMS FOR MOLTEN SALT REACTOR CORE ANALYSIS
2.1. MSBR The Molten Salt Breeder Reactor project by ORNL provides the characteristics of the MSBR reactor concept. The MSBR is a thermal spectrum reactor with a power level of 2250MWth. The MSBR uses molten-salt fuel (LiF-BeF2-UF4-ThF4) and bare graphite moderator. The MSBR is operated by removing fission products and actinides, and adding fertile material (232Th) continuously. It also has characteristics for in dealing with materials during reprocessing & refueling; it can remove all volatile gases (e.g., 135Xe) and noble metals in 20 seconds. The MSBR separates 233 Pa from the molten-salt fuel over 3 days, and allows it to decay to 233U and be recovered. The reactor prevents 233Pa from absorbing neutrons, and increases the efficiency of breeding 233U in the core. Other fission products have specific removal rates. If the depletion time intervals are very short owing to the material being instantly removed, it is hard to calculate long depletion times of MSBR by computer code. For that reason, the unit time interval is set to 3 days stand for the removal time of 233Pa. 232Th is added to breed 233U and to keep the initial amount of 232Th. The total depletion time is 20 years. Figure 1 shows the depletion procedure.
Figure 1. A depletion calculation procedure of the MSBR 2.2. FUJI-U3 The FUJI-U3 was designed by TUT. It is a thermal spectrum reactor with a power level of 450MWth. It uses molten-salt fuel (LiF-BeF2-UF4-ThF4) and a bare graphite moderator. It is operated by removing gaseous fission products and adding fertile (232Th) and fissile (233U) material continuously. It is a converter rather than a breeder. Reprocessing & refueling characteristic is that it usually removes only gaseous fission products and adds 232Th to keep the initial amount of 232 Th and adds 233U so as to have a constant multiplication factor. In addition to the usual material processing, there are chemical processing that make almost the same material composition as the initial operation. Three different chemical reprocessing intervals are conducted in this paper, which are every 2000 effective full-power days (EFPD), 4000 EFPD, and no chemical reprocessing during 8000 EFPD. The gaseous fission products removal ability is 1.4%/min. The primary point of the operation is to maintain the multiplication factor about 1.01 (1.009~1.014). The multiplication factor of the FUJI-U3 is usually maintained by insertion of a graphite control rod or by increasing the fuel flow speed. 233U is added to maintain the multiplication factor every 30 EFPD. For that reason, the time interval is set as 30 days stand for the addition time of 233U. The total depletion time is 8000 EFPD. Figure 2 shows the depletion procedure. PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
3 / 11
Y. Jeong, S. Choi & D. Lee
Figure 2. A depletion calculation procedure of the FUJI-U3 3. PROBLEM SPECIFICATIONS 3.1. MSBR Its reactor core consists of two zones one flow. It means the same composition of molten-salt fuel is divided into two zones in the reactor core, and the graphite to fuel ratio is different in each zone. In this calculation, the two zone design is replaced by a single zone maintaining a volume of molten-salt fuel and graphite. The size of the square shaped graphite cell of Zone I and Zone II are the same. The fuel channel radius is determined by the volume fraction of molten-salt fuel in the unit cell (20.6%) that represents the average volume fraction of molten-salt fuel in Zone I (13%) and Zone II (37%). Figure 3 shows the geometry for the unit cell calculation. Initial fuel loading composition is roughly the same as MSBR, 71.8 LiF – 16 BeF2 – 12 ThF4 – 0.2 UF4. The boundary condition of this unit cell is reflective. The density of the molten-salt is 3.28g/cm3 and that of graphite is 1.84g/cm3 at 900K. The temperature of fuel and graphite is fixed to 900K.
Figure 3. MSBR unit cell geometry for MCNP6 & SCALE/TRITON depletion
4 / 11
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
DEVELOPMENT OF COMPUTER CODE SYSTEMS FOR MOLTEN SALT REACTOR CORE ANALYSIS
3.2. FUJI-U3 Its reactor core consists of three regions one flow. In this calculation, the three region core design is replaced by a single zone, maintaining the composition of molten-salt fuel and graphite. The size of the hexagonal graphite unit cell in each core is the same. The fuel channel radius of a unit cell is determined by the volume fraction of molten-salt fuel in the unit cell (30.8%) that represents the average volume fraction of molten-salt fuel in Core I (34%), Core II (32%) and Core III (26%). Figure 4 shows the geometry for the unit cell calculation. Initial fuel loading composition is roughly the same as FUJI-U3, 71.8 LiF – 16 BeF2 – 12 ThF4 – 0.2 UF4. The boundary condition of this unit cell is reflective. The density of the molten-salt is 3.28g/cm3 and that of graphite is 1.84g/cm3 at 900K. The temperature of fuel and graphite is fixed to 900K.
Figure 4. FUJI-U3 unit cell geometry for MCNP6 & SCALE/TRITON depletion 4. RESULTS Parameters that represent the MSR physics are obtained by the computer code systems. The parameters are change of multiplication factor, breeding ratio, number density of isotopes, amount of total feeding, neutron flux distribution in terms of depletion time, and temperature coefficient. The multiplication factor during depletion is presented by the initial multiplication factor of each time step, because the fission products are produced in computer codes during the time interval. That is because the fission products and fissionable materials are discretely removed and added in the computer code, different from real online reprocessing & refueling. The change of breeding ratio is calculated by the material difference between a small depletion step with an actinide chain and a decay chain of fissile and fertile isotopes. The isothermal temperature coefficient (ITC) is calculated by comparing two multiplication factors; normal temperature, and an increased temperature by 300K for both fuel and moderator.
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
5 / 11
Y. Jeong, S. Choi & D. Lee
4.1. MSBR 4.1.1. Infinite multiplication factor
Figure 5. Multiplication factor of MCNP6 and SCALE for 20 years depletion Figure 5 represents the infinite multiplication factor calculated by MCNP6 and SCALE. The value denoted on the figure is the initial infinite multiplication factor of every depletion time interval. The infinite multiplication factor of MCNP6 is fluctuating because of a high standard deviation (around 200pcm). An infinite multiplication factor calculated with a small standard deviation is the mid-point of fluctuating k because the fluctuating k is bounded by a designated value that is determined by material composition. However k of the two code systems show differences as explained in 4.1.3. 4.1.2. Breeding ratio
Figure 6. Breeding ratio of MCNP6 and SCALE for 20 years depletion In figure 6, the breeding ratio is calculated from the characteristics of isotopes in the thorium cycle and the very fine amount of material information according to short depletion time interval. The average breeding ratio calculated from the result of MCNP6 is 1.0045, and from the result of SCALE is 1.0027. The breeding ratios are slightly higher than one. 6 / 11
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
DEVELOPMENT OF COMPUTER CODE SYSTEMS FOR MOLTEN SALT REACTOR CORE ANALYSIS
4.1.3. Material number density & fuel feeding
Figure 7. Material number density for 20 years depletion of MCNP6 and SCALE Figure 7 shows the number density of major isotopes that have a strong impact on the states of the core. Every isotope represents the beginning of each depletion time interval, except for 233Pa because it is removed in these intervals, so it represents the end of the depletion time intervals. Every isotope denoted in Figure 7, except 235U, is well matched with each other. However, MCNP6 has twice an amount of 235U than SCALE. 235U mainly originates from the 234U neutron capture. It has a large influence on other parameters. Table 1 shows the total amount of 232Th feeding during 20 years depletion. Table 1. Total amount of feeding per unit cell in each computer code system during 20 years Code system Thorium-232 (kg)
MCNP6 4.45
SCALE 4.71
4.1.4. Neutron flux distribution & temperature coefficient
Figure 8. Initial and equilibrium state of normalized neutron flux of MCNP6 and SCALE
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
7 / 11
Y. Jeong, S. Choi & D. Lee
Figure 8 shows the normalized neutron flux distribution of the initial and equilibrium states; from the results, MCNP6 is slightly more moderated in the initial state than SCALE. On the contrary, MCNP6 is slightly less moderate than SCALE in the equilibrium state. The difference of the initial state is negligible, but that of the equilibrium state is considerable. At the equilibrium, MCNP6 has twice the amount of 235U than SCALE, so it consumes a large amount of thermal neutrons. 235U mainly causes the difference in the normalized neutron flux distribution. Table 2 is the isothermal temperature coefficient of the initial and the equilibrium states. The temperature coefficients of MCNP6 and SCALE are similar in the initial state, but they have a large difference not only in value, but also in tendency, in the equilibrium state. Temperature coefficients are mostly affected by the different isotope composition caused by 235U in the equilibrium state.
Table 2. ITC of the initial and equilibrium states of MCNP6 and SCALE. Code system Initial state (pcm/K) Equilibrium state (pcm/K)
MCNP6 -2.53 -4.85
SCALE -2.93 -2.51
4.2. FUJI-U3 4.2.1. Infinite multiplication factor & breeding ratio
Figure 9. Multiplication factor & breeding ratio for different chemical processing interval Three cases being considered in this paper; chemical reprocessing every 2000 EFPD, every 4000 EFPD, and no chemical reprocessing while removing gaseous fission products. Figure 9 shows that the breeding ratio continuously decreases despite the maintained multiplication factor and high breeding ratio made by chemical reprocessing. There is a compatible factor of the three cases. If the reactor operated without chemical reprocessing, then it does not need to pay costs and time, despite the low breeding ratio. On the other hand, the shorter the chemical reprocessing interval is, the higher the average of breeding ratio achieved is, while it needs cost and time. PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future 8 / 11 Kyoto, Japan, September 28 – October 3, 2014
DEVELOPMENT OF COMPUTER CODE SYSTEMS FOR MOLTEN SALT REACTOR CORE ANALYSIS
4.2.2. Material number density & fuel feeding
Figure 10. Material number density of no chemical reprocessing during 8000 EFPD Figure 10 shows that the materials number density reaches equilibrium while operating no chemical reprocessing. During depletion, 232Th is added to maintain its initial amount, 233U is added to keep the multiplication factor as 1.01.Table 3 shows the total amount of fissile and fertile feeding. Fuel feeding increases as chemical reprocessing interval increases.
Table 3. Total amount of feeding per unit cell in each unit cell during 8000 EFPD Time interval between chemical processing Thorium-232 (kg) Uranium-233 (kg)
2000 EFPD 12.41 1.38
4000 EFPD 11.92 1.54
8000 EFPD 11.15 1.92
4.2.3. Neutron flux distribution & temperature coefficient
Figure 11. Initial and equilibrium state of normalized neutron flux of MCNP6 PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
9 / 11
Y. Jeong, S. Choi & D. Lee
Figure 11 shows the MCNP6 calculation of no reprocessing during 8000 EFPD. The neutron flux hardens as fission products are piled up. Table 4 is the isothermal temperature coefficients of the initial and equilibrium states. It shows negative temperature coefficient from initial to equilibrium state and the temperature coefficient gradually decreases.
Table 4. Isothermal temperature coefficient of initial and equilibrium state MCNP6 -2.48 -3.19
Code system Initial state (pcm/K) Equilibrium state (pcm/K)
5. CONCLUSIONS Few contemporary nuclear reactor core analysis codes directly support the modeling of continuous reprocessing and refueling. However, using small discretized depletion time intervals, incorporating with PYTHON script language, makes it possible to model continuous material processing and to obtain various kinds of parameters representing MSRs. The results of MCNP6 (probabilistic method) and NEWT module in SCALE (deterministic method) show some differences in depletion analysis, but it still seems that they can be used to analyze MSRs. In this paper, two MSR core analysis computer code packages have been set up using the two code systems. It has been confirmed that it is possible to analyze various parameters for the MSR unit cells such as the change of multiplication factor, breeding ratio, number density of isotopes, amount of total feeding, neutron flux distribution in terms of depletion time, and temperature coefficient. Furthermore, the two code systems will be able to be used for analyzing whole core models of MSRs. The advantage of the code package using MCNP6 is a high accuracy in transport solutions, but it takes a much longer time to do depletion calculations because of the many short depletion intervals. On the other hand, the advantage of the code package using SCALE is the proven ORIGEN depletion module, but there are differences in the NEWT transport solutions compared with MCNP6. The 234U-235U chain and its data have been identified as one source of the differences in the two implemented code packages for MSR analysis. Further study will be required to address the differences. ACKNOWLEDGMENTS This work was partially supported by National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP). This work was partially supported by the 2013 Future Challenge Research Fund (Project No. 1.130039.01) of UNIST (Ulsan National Institute of Science and Technology).
10 / 11
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
DEVELOPMENT OF COMPUTER CODE SYSTEMS FOR MOLTEN SALT REACTOR CORE ANALYSIS
REFERENCES [1] J. J. Powers, T. J. Harrison, and J. C. Gehin, “A new approach for modeling and analysis of molten salt reactors using scale”, M&C, Sun Valley, Idaho, May 5-9, vol 2, pp. 803 – 815 (2013) [2] “MCNP6 User’s Manual”, LA-CP-13-000634, Version 1.0, Los Alamos National Laboratory Report (2013) [3] “SCALE: A Modular Code System for Performing Standardized Computer Analyses for Licensing Evaluation”, ORNL/TM-2005/39, Version 6, Vols. 1 – 3, Oak Ridge National Laboratory, Oak Ridge, TN, USA (2009) [4] R. C. Robertson et al., Conceptual Design Study of a Single-Fluid Molten Salt Breeder Reactor, Oak Ridge National Laboratory, Springfield, VA, USA, (1971) [5] K. Mitachi, T. Yamamoto, R. Yoshioka, “Three-region core design for 200-MW(electric) molten-salt reactor with thorium-uranium fuel”, Nuclear Technology, 158(3), pp. 348 – 357. (2006)
PHYSOR 2014 – The Role of Reactor Physics Toward a Sustainable Future Kyoto, Japan, September 28 – October 3, 2014
11 / 11
