Conference Paper MCNP5 ANALYSIS OF HISTORICAL CRITICAL SUBSTITUTION EXPERIMENTS IN ZED-2 WITH MIXED TH/235U TEST FUEL COMPANY WIDE CW-115530-CONF-008 Revision 0
Prepared by Rédigé par
Bromley Blair - Reactor Physicist Reviewed by Vérifié par
Chow Jimmy Chun Leun Reactor Physicist Approved by Approuvé par
Radford Darren D. - Manager, Computational Reactor Physics
2010/08/04
2010/08/04
UNRESTRICTED
ILLIMITÉ
Atomic Energy of Canada Limited
Énergie Atomique du Canada Limitée
2251 Speakman Drive Mississauga, Ontario Canada L5K 1B2
2251 rue Speakman Mississauga (Ontario) Canada L5K 1B2
American Nuclear Society (ANS) Annual Winter Meeting. November 7-11, 2010, Riviera Hotel, Las Vegas, NV, U.S.A.
CW-115530-CONF-008
MCNP5 Analysis of Historical Critical Substitution Experiments in ZED-2 with Mixed Th/235U Test Fuel Erik Hagberg, Boris Shukhman, Blair P. Bromley AECL – Chalk River Laboratories, Chalk River, Ontario, ON, Canada K0J 1P0
[email protected],
[email protected],
[email protected] ZED-2 CRITICAL EXPERIMENTS INTRODUCTION Thorium (232Th) has been under investigation and testing since the 1950s [1] as a fertile (and fissionable) fuel for nuclear reactors which could help extend existing uranium resources and possibly replace them in conventional, advanced converter and breeder reactors. Reserves of thorium ore in the earth’s crust are estimated to be two to three times higher than that of the uranium ore. The CANDU (CANada Deuterium Uranium) heavy water reactor is especially well suited for using Th-based fuels because of its excellent neutron economy, which helps minimize the initial feed concentration of fissile isotopes (e.g., 235U, 239Pu, 241Pu, etc.) required to drive the reactor, as the inventory of 233U bred from 232Th builds up. The on-line refueling capability of CANDU reactors with short fuel bundles (~50 cm in length) also allows more flexibility in fuel management in order to maximize fuel burnup and to minimize excess core reactivity. In order to implement Th-based fuel cycles in CANDU reactors, it is necessary to evaluate the accuracy of computer codes and nuclear data libraries that may be used in the reactor physics design and licensing [2]. Therefore, fundamental reactor physics measurements from critical experiments involving Th-based fuels are needed to help better understand the accuracy of physics codes; however, such data are much more limited than that for uranium and plutonium-based fuels. Since the 1960s, Atomic Energy of Canada Ltd. (AECL) has performed extensive research on thorium fuel cycles [3,4,5,6] including a number of critical measurements with a variety of Th-based fuels in substitution experiments [3,4] in the ZED-2 heavy water critical facility at the Chalk River Laboratories [7]. AECL has an extensive database of critical experiments with Th-based test fuels, including U-235/Th, Pu/Th, and U-233/Th. The objective of this work is to describe one of those historical experiments [3], and the results of recent MCNP5 [8] predictions of physics phenomena compared to the measurements.
CANDU (CANadian Deuterium Uranium) is a registered trademark of Atomic Energy of Canada Limited (AECL).
The Th-based substitution experiments chosen for analysis were performed in the ZED-2 heavy water critical facility in 1966 [3]. Reference lattices with 121 aluminum pressure tube / calandria tube assemblies filled with 19-element natural UO2 fuel bundles (19-NU), at hexagonal pitches of 22 cm and 24 cm, and moderated by D2O, were first tested. These were followed by substitution experiments (see Fig. 1) where the central seven channels were replaced with test fuel assemblies. Test Fuel
Reference Fuel
Fig. 1. Layout of Substitution Experiments in ZED-2
19-ThU Test Fuel Assembly
19-NU Reference Fuel Assembly
Fig. 2. Cross Section View of Fuel Assemblies Each reference 19-NU fuel bundle was approximately 49.4 cm in length, and there were five bundles stacked vertically in each lattice site. The 19-element test fuel bundles (19-ThU) used a homogenized mixture of 98.5 wt% ThO2 and 1.5 wt% highly enriched UO2, (93.02 wt% 235U/U), and were cladded with Zircaloy-2, and cooled with either D2O or air (to simulate a void condition). The 19-ThU bundles were approximately 280 cm long and inserted in aluminum coolant tubes, and spanned about the same height as the five vertically stacked reference bundles in adjacent channels. Cross sectional views of the 19-ThU
American Nuclear Society (ANS) Annual Winter Meeting. November 7-11, 2010, Riviera Hotel, Las Vegas, NV, U.S.A. and 19-NU fuel assemblies are shown in Fig. 2, and dimensions are listed in Table I. TABLE I. 19-ThU and 19-NU Specifications Property/Assembly Fuel Fuel density (g/cm3) Pellet radius (cm) Zr-2 clad ID (cm) Zr-2 clad OD (cm) Inner pitch circle radius (cm) Outer pitch circle radius (cm) Al pressure tube ID (cm) Al pressure tube OD (cm) Al calandria tube ID (cm) Al calandira tube OD (cm)
19-ThU (Th,U)O2 9.33 1.153 1.163 1.245 1.468 2.837 7.366 7.620 N/A N/A
19-NU UO2 10.54 1.429 1.438 1.521 1.651 3.188 8.26 8.79 10.16 10.44
In both the reference lattice and substitution experiments, measurements of critical moderator height and other key parameters (moderator purity, temperature, etc.) were taken. The moderator purity and temperature in the substitution experiments ranged from 99.7 to 99.8 at% D2O and 22C to 24C. Measured critical moderator heights in the experiments varied from 176 cm to 203 cm, which included a 15-cm bottom D2O axial reflector. MCNP MODELS A series of MCNP5 models were created to represent the experimental setups in detail. Two basic models were created, one for the reference lattice and the other one for the mixed lattice substitution experiments (see Fig 1). MCNP5, Version 1.4 was used with a cross section library based on ENDF/B-VII.0 [9]. MCNP5 simulations were typically run with more than 100 million neutron histories to ensure that the statistical uncertainty in keff was less than 0.1 mk (10 pcm). SUBSTITUTION ANALYSIS To isolate the properties of the test fuel from that of the reference fuel in the mixed-lattice substitution experiments, substitution analysis was performed using a modified version of MCNP [10,11]. In substitution analysis, a Neutron Production Correction Factor (NPCF) [10,11] is applied to the weight of neutrons born in fission in reference and test fuels to force the MCNP5-calculated value keff to unity to match the conditions of the critical experiment (keff-exp = 1.000). For a lattice of one fuel type (e.g., reference), NPCF = 1/keff. The NPCF is an adjustment factor that effectively changes the value of infinite multiplication factor (kinf) for the lattice. The NPCF for the reference fuel (NPCFref) is determined first from the MCNP5 analysis of the reference experiment. Then, the NPCFref is applied to reference fuel in the mixed-lattice substitution models, and subsequently the NPCF for the test fuel (NPCFtest) is
CW-115530-CONF-008 adjusted until keff-calc =1.000. Then, the isolated value of keff for the test fuel is determined simply by keff-test = 1/NPCFtest. In the hypothetical situation where a critical experiment of pure test fuel was created, the isolated keff-test is what would be calculated by an MCNP simulation, within uncertainties. The use of the NPCF is required in order to isolate the keff for the test fuel from the keff of the reference fuel, given the calculated keff of the mixed-lattice substitution experiment. In both the reference lattice and substitution critical experiments keff-exp=1.000, identically. The MCNP-based substitution analysis method has been tested and validated previously by comparing k eff values determined from both substitution and full-core experiments of test fuel [10,11]. The uncertainties in the derived NPCFtest are based on the propagated MCNP statistical uncertainties in keff, along with experimental uncertainties in moderator purity, temperature, and measured moderator height. In the situation where two mixed-lattice substitution experiments are performed, where the only difference is in the coolant in the test fuel lattice (either with coolant, “cooled”, or without coolant “void”) and the corresponding change in the measured critical moderator height, the keff-test-cool may be different from keff-test-void, which would represent a bias in the calculation of coolant void reactivity (CVR). Through substitution analysis of each experiment, the NPCFtest and the associated uncertainties, NPCFtest, can be isolated to determine the CVR bias and the propagated uncertainties of MCNP5 in the modeling of the test fuel, whereby: CVR = 1/keff-cool – 1/keff-void = NPCFtest-cool – NPCFtest-void. RESULTS Six key experiments were analyzed: two reference lattices with 19-NU at 22 cm and 24 cm hexagonal pitch, and two sets of substitution experiments at each pitch with air or D2O coolant in the test fuel channels containing 19-ThU. Results of the MCNP5 keff calculations for reference lattices and mixed lattice substitution experiments are shown in Table II. The bias in keff ranged from -3.5 mk to -5.1 mk, and may be attributed to approximations made in the modeling of the experiments, along with potential errors remaining in the nuclear data. Through substitution analysis, NPCFtest ranged between 0.999 and 1.001. Thus, the 19-ThU had an isolated keff bias that ranged from approximately -1 mk to +1 mk. The propagated uncertainties in NPCFtest were approximately 2.1 mk or less. Thus, the MCNP5 calculation bias in the estimate for the coolant void reactivity (CVR) was found to be approximately +0.2 mk to +0.6 mk, with a combined uncertainty of less than 2.8 mk. These results demonstrate that MCNP5 predicts a very low bias in the estimate of CVR for the 19-ThU test fuel.
American Nuclear Society (ANS) Annual Winter Meeting. November 7-11, 2010, Riviera Hotel, Las Vegas, NV, U.S.A. SUMMARY MCNP5 with the ENDF/B-VII.0 nuclear data library has been used to analyze critical mixed-lattice substitution experiments in the ZED-2 heavy water critical facility with thorium-based test fuel. Calculations have shown the bias in keff to range from -1 mk to +1 mk, and bias in coolant void reactivity to range from +0.2 mk to +0.6 mk with an propagated uncertainty of less than 2.8 mk. ACKNOWLEDGMENTS Appreciation is extended to Bruce Wilkin, Choy Wong, Michael Zeller, Greg Morin, David Watts, Alex Trottier, (AECL, Chalk River Laboratories), and the late Albert Okazaki for their insights and assistance with this work. REFERENCES 1. T. PIGFORD, et al., “Fuel Cycles in Single-Region Thermal Power Reactors”, Proceedings of the 2nd United Nations International Conference on the Peaceful Uses of Atomic Energy, P/1016, Volume 13, pp. 198-228, (September 1958). 2. P. G. BOCZAR, et al., “Qualification of Reactor Physics Toolset for a Thorium-Fuelled CANDU Reactor”, Proceedings of 18th International Conference on Nuclear Engineering, Xi’an, China, (May 2010). 3. A. OKAZAKI and S.A. DURRANI, “Lattice Experiments with 19-Element Rods of ThO2 / 235UO2 in Heavy Water Moderator”, AECL-2778, (1967).
CW-115530-CONF-008 4. M.B. ZELLER et al., “A Comparison of Calculations with WIMS-AECL to Thorium Fuel Measurements in ZED-2”, Proceedings of the. 12th Annual Conference of the Canadian Nuclear Society, Saskatoon, Saskatchewan, (June 1991). 5. H. Hamilton and S. Livingstone, “Thorium Fuel Fabrication and Testing at AECL”, Proceedings of the International Workshop on Thorium Utilization for Sustainable Development of Nuclear Energy – TU2001, Beijing, China, (December 2007). 6. P. G. BOCZAR, et al., “Thorium Fuel Cycle Studies for CANDU Reactors,” Thorium Fuel Utilization: Options and Trends, IAEA TECDOC 1319 (2002). 7. K. SERDULA, “Lattice Measurements with 28-Element Natural UO2 Fuel Assemblies Part I: Bucklings for a Range of Spacings with Three Coolants”, AECL-2606, AECL Chalk River Laboratories, (July 1966). 8. X-5 MONTE CARLO TEAM, “MCNP – A General Monte Carlo N-Particle Transport Code, Version 5”, LA-UR-03-1987, (April 2003). 9. D.V. ALTIPARMAKOV, “ENDF/B-VII.0 versus ENDF/B-VI.8 in CANDU Calculations”, Proceedings of the PHYSOR 2010 Conference, Pittsburgh, PA, (May 2010). 10. B.P. BROMLEY et al., “Development and Testing of a MCNP-Based Method for the Analysis of Substitution Experiments”, Transactions of the ANS, Volume 97, pp. 711-712, (November 2007). 11. Y. DWEIRI et al., “Comparison of Bare-Lattice Calculations Using MCNP Against Measurements with CANFLEX-LEU in ZED-2”, Proceedings of the 29th Annual Conference of the Canadian Nuclear Society, Toronto, ON, Canada, (June 2008).
TABLE II. MCNP5 Results for ZED-2 Including Substitution Analysis Pitch (cm) 22 22 22 24 24 24
Ref. Fuel 19-U 19-U 19-U 19-U 19-U 19-U
Test Fuel 19-U 19-ThU 19-ThU 19-U 19-ThU 19-ThU
Test Coolant D2O D2O Air D2O D2O Air
keff 0.99527 0.99646 0.99627 0.99487 0.99597 0.99605
NPCF Ref. 1.00475 1.00475 1.00475 1.00502 1.00502 1.00502
NPCF Test n/a 0.99961 0.99942 n/a 1.00089 1.00028
NPCF Test (mk) 0.2 1.7 2.0 0.2 2.1 1.8
CVR Bias (mk) n/a +0.2 2.6 n/a +0.6 2.8
