Download presentation
Presentation is loading. Please wait.
Published byAllison Norris Modified over 9 years ago
1
Methodology to address radioprotection and safety issues in the IFMIF/EVEDA accelerator prototype J. Sanz 1,2, P. Sauvan 1,2 F. Ogando 1,2, D. López 1,2, M. García 1,2, A. Mayoral 1,2, F. Ortiz 1 A. Ibarra 3, V. Blideanu 4, P. Joyer 4 (1)Departamento de Ingeniería Energética, Escuela Técnica Superior de Ingenieros Industriales, Universidad Nacional de Educación a Distancia (UNED), C/ Juan del Rosal 12, 28040 Madrid (2) Instituto de Fusión nuclear, Universidad Politécnica de Madrid (UPM), C/ José Gutiérrez Abascal 2, 28006 (Madrid) (3) CIEMAT, Madrid, España (4) Commissariat à l’Energie Atomique, CEA/IRFU, Centre de Saclay, 91191 Gif sur Yvette cedex, France Shielding Aspects of Accelerators, Targets and Irradiation Facilities (SATIF-10) Geneva, Switzerland, June 2–4, 2010,
2
Radioprotection and safety issues in the IFMIF/EVEDA accelerator prototype General Computational Methodology Flowchart: issues to discuss Testing available deuteron cross sections: Benchmarking against experimental EXFOR data * Validation of nuclear models included in MC transport codes: MCNPX,PHITS * Interest of the TALYS nuclear model code to simulate deuteron reactions MCUNED: new capabilities for Monte Carlo simulation of light ions transport and secondary products generation * New capabilities: extensions to MCNPX * Verification of MCUNED capabilities MCUNED applications to the validation process of TENDL for deuterons: Analysis of integral experiments for neutron production Proposals under discussion to deal with other RP issues Summary Overview of presentation
3
IFMIF/EVEDA prototype accelerator (Project under Fusion Broader Approach Japan/EU) IFMIF: International Fusion Materials Irradiation Facility 2 deuteron beams (125 mA X 2) up to 40 MeV EVEDA phase: Engineering Validation and Engineering Design Activities 1 deuteron beam (125 mA) up to 9 MeV BD 100 keV 165 mA 5 MeV 140 A 9 MeV 125 mA Faraday Cup normal incidence Collimator in normal operation or interceptive diag at low d.c. (10 -3 ) normal incidence Beam Dump angle of incidence according to design Loss in RFQ ~ 10 mA max DTL and HEBT 1 W/m ~ 100 nA/m
4
Potential radioprotection issues for the IFMIF/EVEDA prototype accelerator Some radioprotection issues Beam-on phase: * Generation of neutrons and gammas by d-target material and d-D interactions: dose levels inside and outside the vault * Tritium production Beam-off Phase: * Deuteron and neutron induced activation: Residual gamma dose rates BD cartridge: Definition of local shielding M. García Talk Concrete walls of the accelerator vault: Thickness = 1.5 meters CONCRETE SHIELD COMPOSITION 0.56 % H 2.21 g/cm3
5
General Computational Methodology: calculational tools Particle TRANSPORT (MC codes) Activation System Neutron and deuteron fluxes Gamma TRANSPORT Isotopic inventory Gamma Source Gamma Dose Neutron Source d-D Neutron Source d-Cu Beam dynamics code Module for neutron Source modeling Module for neutron Source modeling Module for Deuteron source modeling BEAM OFF Neutron and gamma Fluxes Neutron and gamma Doses Dose conversion factors BEAM ON
6
Deuteron transport, interaction and secondary particle production * Monte Carlo transport codes: built-in semi analytical models to describe nuclear interactions of deuterons with mater. - For incident deuterons of energy relevant to IFMIF/EVEDA accelerator: * are the models reliable enough? * can computing time consumption be a concern? Nuclear reaction codes applications to calculate deuteron cros sections Explore de interest for generation of evaluated XS libraries by dedicated nuclear model codes: TALYS code and TENDL library Validation of deuteron nuclear data Validation (first step: basic experimentally measured nuclear data; EXFOR) Validation (second step: integral experimentally measured data) Reliability of existing methodologies: Availability, development a validation of deuteron nuclear data
7
MC transport code LIMITATIONS: MCNPX and others (PHITS) Elements studied : Typical of accelerator equipments Cu, Ni, Fe, Cr, Nb and W ENERGY RANGE : 0 to 20 MeV NO nuclear models in MCNPX ALLOW A GOOD FIT WITH EXPERIMENTAL VALUES (in shape and amplitude). Strong disagreement is observed for main reactions Relevant impact on the prediction of the secondary neutron source production
8
In a few cases MCNPX cross sections are in reasonable agreement with experiments, but with an unreliable spectrum of emitted neutrons MCNPX generate unphysical high energy neutron tails above the maximum physical energy (about 15.5 MeV for Cu63) Big impact on neutron activation (production of radionuclides of concern from neutron threshold reactions) MC transport code LIMITATIONS: MCNPX and others (PHITS)
9
Spectrum of the emitted neutrons is unphysical: high energy tail is computed MC CODES LIMITATIONS: NEUTRON EMISSION Natural Cu
10
WHAT CAN THE TALYS CODE (and TENDL) OFFER? TALYS (global OM parameters) gives a shape that fits rather well the shape of experimental values TENDL=TALYS based library However, the results are in general not good enough in amplitude (not much better than in MCNPX in several cases) Potential to change Optical Model parameter in order to explore the possibility to obtain a right fit to experimental data
11
Even in those cases in which TALYS results are not good enough in amplitude, neutron spectrum is consistent with energy conservation POTENTIAL OF TALYS CODE (and TENDL)
12
Spectrum of the emitted neutrons is consistent with energy conservation TALYS-TENDL: NEUTRON EMISSION Natural Cu
13
USE OF APPROPRIATE ADJUSTING PARAMETERS CAN ALLOW A GOOD REPRODUCTION OF EXPERIMENTAL CROSS SECTIONS? The fitting set of parameters has to be defined for each element. Figures show the improvement obtained with the fitting set of parameter: 56Fe (UNED) 65Cu (Avrigeanu)
14
One of the best know transport code is MCNPX They have some problems for IFMIF-EVEDA use: 1) NO nuclear models in MCNPX ALLOW prediction of excitation functions showing A GOOD FIT WITH EXPERIMENTAL VALUES for interactions of D with typical elements of accelerator equipments in the energy range of interest 2) MCNPX provide unreliable spectrum of emitted neutrons * unphysical high energy neutron tails above the maximum physical energy * for some models not able to provide neutrons for deuteron incident energies below 5-6 MeV 3) Computing time consumption concerns for the low neutron yields in the EVEDA scenario Similar problems can be found in other standard transport codes (for example PHITS or FLUKA) Why extensions to current MC transport codes ? External light ions evaluated libraries (for example TENDL – generated from TALYS code-) must be used We need a code able to read external libraries McUNED
15
MCUNED features and capabilities MCUNED is a MCNPX extension allowing handle light ions evaluated libraries and incorporating a variance reduction technique for the production of secondary particles by light ions induced nuclear reactions Able to handle proton, deuteron, triton, helium-3, alpha evaluated nuclear data libraries for nuclear reactions in transport calculations. Include a variance reduction technique for the production of secondary particles. Save drastically the computing time. Maintain all MCNPX capabilities : – Same input as MCNPX (fully compatible with MCNPX input) – Same particle transport algorithm – Same flexibility in source and geometry definition – Same calculation capabilities (Flux, dose, mesh etc… tallies)
16
Verification of handling light ion libraries 9 MeV deuteron Cu63 target 0.1 m thickness Thin Target experiment on Copper 63 Neutron spectrum produced by the simulation are recorded at 4 different angles They are compared with double differential XS data used in the calculations ( XS data from Tendl 2008 are extracted with Janis 3.0 converted into laboratory system) Satisfactory results
17
Verification of Variance Reduction Technique The verification of the Variance Reduction technique is carried out by comparing MCNPX and MCUNED results with incident proton simulations. Incident proton simulation are used since MCNPX could use only proton data library The same input file is used in both simulation The number of initial proton histories in both simulations is adjusted to achieved the same statistical error in the tally value. The Variance Reduction technique is considered correct if the relative error between MCNPX and MCUNED calculation is comparable to the statistical error of the tally value Relative error Statistical error
18
The same simulation with proton on thick target have been performed by MCNPX and MCUNED 10 MeV proton Natural Copper target 1 mm thickness The proton history number have been chosen in both cases to reach the same statistical error (about 1%) The agreement between both simulations is very good Verification of Variance Reduction Technique
19
Verification in Response function calculations Neutron ambient equivalent dose calculated for a 100 mA proton beam @6.7 MeV on cylindrical thick Nickel Target MCNPX simulation Neutron dose map MCNPX and MCUNED simulations are performed for dose calculation Both results have the same statistical error
20
MCNPX Simulation Statistical error map MCNPX,MCUNED Relative error map Relative errors between MCNPX and MCUNED mean values are lower than the statistical errors of the given response function Spatially averaged statistical error is about 4% Verification in Response function calculations
21
Performance of the Variance Reduction The reduction variance technique is useful when the production of secondary particle is low (it is essential when the production rate of secondary particle is very low) Number of secondary particle per primary particle T p Time needed to transport the primary particle T s Time needed to transport the secondary particle Summary of 10 MeV Proton on Copper simulation In this case the time gain is 4000 !!!
22
TENDL Library TENDL : Talys-based Evaluated Nuclear Data Library Library generated by talys nuclear code Available libraries for incident neutron, proton, deuteron, triton, helion, alpha Incident energy up to 200 MeV Data available for 1000 isotopes Libraries available in ENDF and ACE format http://www.talys.eu/tendl-2009/
23
Checking TENDL library against integral experiments [1] L. W. Smith and P. G. Kruger, “Thick Target Yields from (d,n) Reaction at 10 MeV”, Physical Review, 83, 1137, (1951). [2] A. J. Allen, etal., “Thick Target Fast Neutrons Yield from 15-MeV Deuteron and 30-MeV Alpha-Bombardement ”, Physical Review, 81, 536, (1951). [3] K. Shin etal., “Neutron and photon production fron thick targets bombarded by 30-MeV p, 33-MeV d, 65-MeV 3He, and 65-MeV a ions: Experiment and comparison with cascade Monte Carlo calculations”, Physical Review C, 29, 1307, (1984). [4] J. P. Meulders, etal., “Fast Neutron Yields and Spectra from Targets of Varying Atomic Number Bombarded with Deuterons from 16 to 50 MeV”, Phys. Med. Biol., 2, 235, (1975). Total neutron Yield (4 ) Ref E d (MeV) E n > (MeV) Exp (n/d) TENDL (n/d) Yield frac. 11008.81E-46.44E-41 23301.81E-21.50E-21 Ref Ed (MeV) E n > (MeV) Exp (n/d/Sr) TENDL (n/d/Sr) Yield frac. 31642.76E-45.94E-50.20 23346.16E-37.49E-40.39 33346.68E-37.49E-40.39 Neutron Yield in forward direction Total yields from simulations are in reasonable agreement with experimental values TENDL underestimates neutrons emitted in the forward direction Deuteron induced neutron production in a thick copper target Spectra of neutrons emitted in the forward direction
24
Checking TENDL library against integral experiments Normalized neutron angular distribution Thick copper target bombarded by deuterons Forward neutron spectra are not reproduced by TENDL Experimental angular distribution exhibit a forward peaked neutron emission not reproduced by TENDL. Differences become lower for lower incident energies Backward neutron spectra from simulation are in good agreement with experiments
25
Checking TENDL library against integral experiments Thick target experiment with deuteron on Aluminum Reasonably agreement with Tendl at low deuteron energy T.N.Massey etal. “A Measurement of the 27Al(d,n) Spectrum for Use in Neutron Detector Calibration” Nuclear Science and Engineering, Vol.129, p.175 (1998) M.Hagiwara etal. “Experimental studies on the neutron emission spectrum and activation cross-section for 40 MeV deuterons in IFMIF accelerator structural elements” Journal of Nuclear Materials, Vol.329, p.218 (2004) Aluminum 6mm thick, Ed=40 MeVThick Aluminum foil, Ed=7.44 MeV
26
Issue: Deuterium Concentration Profile inside the Copper Lattice and n production Some Proposals under discussion Deuterium implantation profile: MCNPX/MCUNED, SRIM Deuteron concentration profile: New transport coefficient data for IFMIF/EVEDA operation conditions and TMAP code calculations Neutron production: MCUNED. OTHER ELEMENTS OF THE COMPUTATIONAL METHODOLOGY (still under discussion within the Project) Issue: Deuteron, proton a neutron induced activation Some Proposals under discussion ACAB activation code. EAF data libraries Issue: Tritium production by deuteron induced reactions (d-D, d-Cu) Some Proposals under discussion Tritium production and implantation profile: MCUNED Tritium concentration in structural materials, and diffusion to water, vacuum pump: Transport coefficient data for IFMIF/EVEDA operation conditions and TMAP calculations.
27
Summary Different alternatives have been found regarding the different computational elements needed in addressing the RP issues of the EVEDA accelerator prototype. Some alternatives are under discussion within ASG between CEA and UNED regarding activation and deuterium implantation issues. Final agreement will be reached soon Regarding simulation of deuteron transport and secondary products generation the agreement is already reached 1.The nuclear reaction models included in the transport codes (MCNPX, PHITS) cannot be used in the EVEDA RP calculations. 2.The cross-sections of the reactions should be re-assessed with the TALYS code which allows a better description of the nuclear reactions of interest for EVEDA 3.Extension of Monte Carlo codes to use evaluated data files generated by TALYS with appropriate adjusting OMP parameters 4.MCUNED code, a MCNPX extension is the current response to this need.
28
Summary MCUNED: Two extensions to MCNPX * Handling light ion data libraries (energy-angle distributions of all outgoing particles): enables to include updated/reliable nuclear cross section for transport simulation * Reduction variance technique for generation of secondary products: drastic reduction in the computing time needed for a target accuracy MCUNED has been verified with very positive results Valuable in radioprotection studies on accelerator design: it is specially targeted for low-energy light ion applications Superior performance of MCUNED-TENDL libraries versus current Monte Carlo simulation tools for deuterons in the EVEDA energy range: adjusting D-TENDL library required? Useful tool for Benchmarking evaluated cross section against integral experiment
29
Summary Efforts are underway to evaluate TENDL for incident deuterons on copper against integral experiment for neutron production. Total neutron yields from simulations are in a reasonable agreement with experiments Angular distribution and spectrum of backward-emitted neutron is well reproduced by TENDL TENDL underestimates neutrons emitted in the forward direction Neutron spectra not reproduced for neutrons emitted in the forward direction Differences between experiments and simulations decreases as incident deuteron energy decreases, and a reasonable behavior of TENDL library could be expected for EVEDA applications Few experiments exist for EVEDA applications. Very recent experiments will be very useful in evaluating and improving (if necessary) TENDL for EVEDA applications.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.