Automated calculation sequence for group constant generation in Serpent 4th International Serpent UGM, Cambridge, UK, Sept. 17-19, 014 Jaakko Leppänen VTT Technical Research Center of Finland
Click to edit Master title Outline style What is spatial homogenization? Covering Second all level state points Homogenization methods in Serpent Automated Third calculation level sequence in Serpent Fourth level Example case: Serpent-ARES code sequence Fifth level Estimation of computational cost Summary and conclusions
Second level Transport problem becomes non-linear when feedbacks from material temperatures Third level and densities are taken into account Same applies Fourth to changes level in fuel composition with increasing burnup What is spatial homogenization? Modeling of an operating nuclear reactor is a complicated task: Full-scale solution to neutron transport problem is computationally expensive Fifth level In practice, the solution to the coupled problem is obtained by iteration between: Solution of linearized transport problem Solution of heat transfer and coolant flow Solution of Bateman depletion equations There exists high-fidelity solution methods for the independent problems (e.g. continuousenergy Monte Carlo and CFD), which can be combined for the solution of the coupled problem, but the approach is not practical for routine design and safety analyses
Second level Third level Fourth level Fifth level What is spatial homogenization? Instead, the solution is obtained by gradually increasing the scale of the system, while simplifying the physics. The calculation sequence essentially consists of two parts: i) Spatial homogenization: Interaction physics at the fuel assembly level is condensed into a set of representative group constants Geometry is homogenized, energy dependence condensed into few energy groups Local reaction rate balance is preserved Traditionally based on D transport calculation ii) Full-core calculation: Group constants from spatial homogenization are used as the building blocks for a simplified full-core calculation In LWR applications typically based two-group nodal diffusion methods Neutronics solution is obtained at an acceptable computational cost, which enables iterative solution to the coupled problem
Second level Third level What is spatial homogenization? Traditional methods for spatial homogenization: Deterministic transport solution, for example, MOC or CP Two-dimensional infinite-lattice geometry Advantages of using Fourth continuous-energy level Monte Carlo for homogenization: Fifth level No major approximation for geometry or physics Straightforward, no intermediate solutions to account for self-shielding effects Inherently three-dimensional Same code and cross section data can be used for any fuel or reactor type Serpent is one of the first Monte Carlo codes designed from the beginning for the purpose of spatial homogenization
Second level Scattering Third matrices level Diffusion coefficients Fourth level Homogenization methods in Serpent Serpent has the capability to produce homogenized multi-group constants for nodal diffusion reactor simulator calculations: Homogenized reaction cross sections Fifth level Assembly discontinuity factors Form factors for pin-power reconstruction Methods exist for the calculation of effective delayed neutron fractions, but the routines are not specifically designed for homogenization: Based on the iterated fission probability (IFP) method Calculation covers the entire geometry (not applicable to assembly colorsets)
Homogenization methods in Serpent The calculation is based on the preservation of local reaction rate balance and two built-in deterministic solvers: i) B 1 critical spectrum calculation: Second level Homogenized reaction cross sections are calculated using an intermediate micro-group Third level structure (by default WIMS 69-group structure) Fourth level B 1 -equations are formed and solved by critical buckling iteration 1 Fifth level The result is a leakage-corrected micro-group spectrum, which is used for collapsing the cross sections into group constants using the final macro-group structure (by default energy groups) ii) Homogeneous diffusion flux solver for ADF s and form factors: Based on the solution of two-dimensional diffusion equation in the homogenized region, using net currents as boundary conditions Used when homogenization is performed for assembly colorset or reflector (non-zero net current) More about this in Maria s presentation 1 E. Fridman and J. Leppänen. On the Use of the Serpent Monte Carlo Code for Few-Group Cross Section Generation. Ann. Nucl. Energy 38 (011), 1399 1405.
Homogenization methods in Serpent The methodology has been tested with different code sequences and reactor types: Second level Serpent-DYN3D: VVER-440, HTGR 3, SFR 4 Serpent-PARCS: PWR 5, SFR 6, RBWR 7 Serpent-ARES: Third level MIT BEAVRS Benchmark (PWR) 8 Fourth level Fifth level This is not the complete list more examples at this meeting and PHYSOR 014 S. Duerigen and E. Fridman. The Simplified P3 Approach on a Trigonal Geometry of the Nodal Reactor Code DYN3D. Kerntechnik 77 (01), 6 9. 3 S. Baier et al. Extension and application of the reactor dynamics code DYN3D for Block-type High Temperature Reactors. Nucl. Eng. Design 71 (014), 431 436. 4 E. Fridman and E. Shwageraus. Modeling of SFR Cores With Serpent-DYN3D Codes Sequence. Ann. Nucl. Energy 53 (013), 354 363. 5 M. Hursin et al. Comparison of Serpent and CASMO-5M for Pressurized Water Reactors Models. In proc. M&C 013. Sun Valley, ID, 013. 6 L. Ghasabyan. Use of Serpent Monte Carlo Code for Development of 3D Full-Core Models of Gen-IV Fast-Spectrum Reactors and Preparation of Group Constants for Transient Analyses with PARCS/TRACE Coupled System. M.Sc. Thesis, Royal Institute of Technology. 013. 7 A. Hall et al. Advanced Neutronics Methods for Analysis of the RBWR-AC. Trans. Am. Nucl. Soc. 108 (013), 771 774. 8 J. Leppänen, R. Mattila, and M. Pusa. Validation of the Serpent-ARES Code Sequence Using the MIT BEAVRS Benchmark Initial Core at HZP Conditions. Ann. Nucl. Energy 69 (014), 1 5.
Homogenization Click to edit Master titlemethods style in Serpent R P N M 1 3.08 L K J H G F E 1.7 1.53 1.64 1.74 1.53 1.7 1.5 1.6 1.89 1.88 1.96.07 1.70 1.7 1.99.06 1.70 1.7 1.99.05 1.50 1.96 1.89 1.50.06 1.94 1.88.05 1.99 1.7 1.70.05 1.99 1.69 1.97 D C B 6 1.98 1.99.41.41 1.7 1.70 1.49 1.7 1.86.4.41.00 1.99.09.06 1.95 1.75 1.50 1.68 1.50 1.76 1.86 1.7 1.97 1.7 1.96 1.7 1.74 1.66 1.74 1.94 1.5 8 9 11 4 1.75 1.49 1.50 1.76 1.70 1.74 1.61 1.61 1.74 1.88 1.49 1.33 1.40 1.34 1.50 1.7 1.88 1.61 1.33 1.33 1.61 1.70 1.88 1.66 1.3 1.40 1.68 1.49 1.64 1.7 1.87 1.88 1.69 1.89 1.66 1.4 1.66 1.70 1.49 1.61 1.33 1.33 1.33 1.33 1.61 1.49 1.7 1.89 1.88 1.69 1.74 1.49 1.61 1.49 1.75 1.70 1.53 1.69 1.74 1.61 1.61 1.74 1.70 1.95 1.84 1.61 1.75 1.49 1.66 1.49 1.75 1.7 1.96 1.6 K 0.0 - -1. - - - - - -0.9-1.5 J - H G -0.7 F - E D C B A -1.1 0.0 0.0 - -0.9-0.7-1.0 - -1.0 5 - -0.8-0.9-1.1-0.9-1.0-1.4-1.6 6 0.0-0.8-1.0-1.0-1.1-0.7-1.0-0.9-1.0 7 0.0 - -0.7 - - -0.8-1.0-1.1-1.0-0.7 1 1.74.07 1.66 1.74 1.7 1.96.41 1.99.07.06 1.99.41 1.70.41 1.99.07.04 1.98 1.69 1.70 1.98.05 1.50 1.98.05.41 1.99 1.7 1.70 1.69 1.99.05.07.07.07.06 1.50 14 1.95 1.96 1.87 1.88 1.88 1.88 1.96 1.94 1.6 1.53 1.6 1.6 1.5 Fifth level 15 P 8 - - -1.1-0.7 - - - -0.7-0.8-10 -0.0 0.0 - - - - -0.7-0.8 - - -0.0-11 - - - - - - 0.0-1 - - - - - - 13-0.9 0.5-0.0-0.0 - - - -1. D C 9 1 0.6 - -0.0 - - -0.8 - -0.7 0 1 1.94 Fourth level R - 1.74 1.6 1.5 L 13-0.9-1.54 1 M 1.5 Third level 10 1.69-0.0 3.07.07 N -1.0 - -1.1 3 1.6 Second level 7 P -0.7.5 1.6 4 R 1.53 Click to edit Master text styles 5 A 1.08 N M L K 1.6 0.5 14 15 - - - - - -0.8 F E 0 J H G F E D C B 3 A R P N M L K J H G B A 3 1 1.5 3 3 4 5 4 5 1 6 6 7 7 1.5 8 8 9 10 0 9 1 10 1 11 11 1 1 13 0.5 13 14 14 15 15 0 3 Figure 1: Results of Serpent-ARES calculations compared to Serpent 3D reference results: assembly (top row) and pin level (bottom row) power distributions (left column) and relative differences (right column).
Most of the previous studies were limited to hot zero-power conditions and initial cores. In reality, spatial homogenization is complicated by the fact that group constants de- pend Click on: to edit Master text styles Second level Local operating conditions: fuel temperature, moderator temperature (PWR), coolant Third void level fraction (BWR), boron concentration (PWR) Fourth level Fuel burnup Fifth level Covering all state points Operating history: local void fraction (BWR) or moderator temperature (PWR), boron concentration (PWR), position of control rods (BWR) In order to provide the sufficient building blocks for the full-scale simulation, spatial homogenization must cover the full scope of reactor operating conditions
Second Serpent-PARCS level code sequence 9 Covering all state points Covering all assembly types, burnup points and operation condition requires hundreds or thousands of runs. There exists codes and scripts for automating the task: SerpentXS Python-based wrapper script developed by Bryan Herman from MIT for SP Third Matlab-based level wrapper script developed by Levon Ghasabyan from PSI for Serpent-PARCS Fourth level code sequence 10 Fifth level PyNE Python-based utility code package developed by Anthony Scopatz from University of Wisconsin-Madison 11 GenPMAXS The preprocessor code for PARCS (the latest version has Serpent support) Problem: All these scripts were developed for Serpent 1, and they are not compatible with the new variables in Serpent (according to my knowledge). 9 http://canes.github.io/serpentxs/index.html 10 https://github.com/tumregels/stop 11 https://github.com/pyne/pyne
Automated calculation sequence in Serpent To simplify group constant generation, an automated calculation sequence is currently being developed for Serpent : Second level Based on restart calculations, performed after the main burnup cycle Capable Third of handling level branches to different state points Writes a separate Fourth level output file, easily read to processing scripts Fifth level Covers only branch calculations, history calculations must be run separately Available since version.1.1, important updates in.1.. The work is still under way, and there is a separate topic for discussion at the Serpent forum: http://ttuki.vtt.fi/serpent
Automated calculation sequence in Serpent The calculation sequence is based on branches ( branch card), used to inflict changes into original input file: Second level Change in material density and temperature ( stp entry) 1 Replace Third onelevel material with another ( repm entry) Replace one Fourth universe level with another ( repu entry) Fifth level And a coefficient matrix ( coef card): Defines burnup points for which branch calculations are performed Defines combinations of branches run for each burnup point 1 The code automatically retrieves the correct cross sections from the directory file and performs Doppler-broadening on the data. Variations in moderator temperature can be inflicted by replacing the entire material.
% - Nominal state branch (do nothing): branch nom Second level Third level Fourth level Simplified input example % - Fuel temperature branches (change in material temperature): branch fueh stp fuel4-157 1600 branch fuec stp fuel4-157 65 % - Coolant boron concentration branches (replace material): Fifth level branch borl repm cool cool_lob branch borh repm cool cool_hib % - Control rod insertion branch (replace universe): branch CR repu T R % - Coefficient matrix (13 burnup points, 3x3x branch combinations): coef 13 0 5 10 15 0 5 30 35 40 45 50 55 60 3 nom fueh fuec 3 nom borl borh nom CR
Example case: Serpent-ARES code sequence Some preliminary tests have been carried out for the Serpent-ARES code sequence. Observations: Second level It was possible to define the branches in such way that all state points were covered, but the calculations resulted in several combinations that were not used for anything Third level Fourth level A post-processing Fifth script level was relatively easy to write for the output file, but the format had to be changed a few times along the way the format is likely to change in near future updates as well Use of variables was a convenient way to pass history information into the processing script The methods were also successfully applied to HZP initial core calculations with Serpent- TRAB3D (a transient code developed at VTT)
Example case: Serpent-ARES code sequence ARES reads group constants for each control rod type, burnup and void (BWR) or moderator temperature (PWR) history and the values for the local homogeneous flux solution inside each node are obtained by table interpolation and polynomial fit: Second level Σ = Third Σ 0 + level (aρ + b)ρ + (ct f + d)t f + (et m + f )T m + gρt f Fourth + (htlevel m + i)ρt m + jt f T m + kh cr + (lb + m)b Fifth level + (nh bo + o)h bo + (pb Tm + qb T m + vbtm + wbt m) + (rb + s)bt f + (tb + u)bρ, where Σ 0 is the value at the nominal state, a... w are the polynomial coefficients and ρ is the relative coolant density - nominal value T f is the square root of fuel temperature - nominal value T m is the moderator temperature H cr is the control rod history (BWR) B is the boron concentration (PWR) H bo is the boron history (PWR)
Example case: Serpent-ARES code sequence The previous HZP initial core BEAVRS 13 benchmark calculations were repeated with cross sections parametrized for boron concentration: Second level Good results for radial and axial power distribution compared the 3D Serpent reference calculation Third level Critical boron Fourth concentrations level and control rod bank worths consistent with benchmark values Fifth level The next stage is to proceed to HFP and burnup: Critical boron for HFP state with equilibrium xenon consistent with benchmark value Comparison to Serpent 3D calculation under way (fuel temperature and moderator density/temperature from ARES to Serpent reference calculation) Processing script is still missing the calculation of cross terms 13 http://crpg.mit.edu/pub/beavrs
Example case: Serpent-ARES code sequence Second level Third level Table 1: Critical boron concentrations (ppm). Configuration ARES Ref. A-R ARO 989 975 14 D 96 90 4 C,D 84 810 14 A,B,C,D 683 686-3 Fourth level A,B,C,D,SE,SD,SC 486 508 - Fifth level Table : Control rod bank worths (pcm). Bank Serpent 3D ARES Ref. A-S A-R D 781 (5) 787 788 6-1 C 144 (5) 145 103 1 4 B 1199 (6) 1197 1171-6 A 538 (6) 58 548-10 -0 SE 497 (5) 497 461 0 36 SD 785 (6) 783 77-11 SC 110 (6) 1093 1099-7 -6
two Click questions: to edit Master text styles Second level Estimation of computational cost Monte Carlo simulation is computationally expensive compared to deterministic transport methods, and the results are always accompanied by statistical error. This raises i) How many neutron histories need to be run for spatial homogenization to reduce the statistical Third level errors to such level that random noise is not reflected in the simulator calculation? Fourth level Fifth level ii) What is the overall computational cost for generating the full set of group constants, covering all assembly types, burnups and state points? These questions are addressed in a paper presented at PHYSOR 014 in two weeks, 14 using the BEAVRS benchmark as a test case. 14 J. Leppänen and R. Mattila. On the Practical Feasibility of the Continuous-energy Monte Carlo Method for Spatial Homogenization. In proc. PHYSOR 014. Kyoto, Japan, Sept. 8 - Oct. 3, 013.
Estimation of computational cost.0 4 0.8 Click to edit Master.9 0.7 1. 0.5 0.7 5 6 text 1..7 0.6. 0.7.9 0.6 styles 1.6 1. 1.5 0.5 1.8 0.7.1 0.8.5 0.7.1 0.7 1.3 0.5 1.1 0.5 3.0 0.7 3.1 1.0 4.0 1.0 3. 1.0 3.4 0.7 1.7 0.5 1.3 0.6 1.5 0.7 1. 0.7 3.9 0.8.7 0.5 0.8 0.9.9 0.7 3.1 1.1 4.5 1. 4.0 1. 4.7 1. 3.7 1.1 3.5 0.7 1.7 1.4 1.6 0.5. 0.8 3.3 7 0.6 1.6 0.5 1.0.0 0.6.8 1.0 4.3 1. 4.1 1.3 4.7 1.1 4.5 1.3 4.5 1. 3.4 1.0 3.0 0.6 1.0 1.7 0.5.0 0.6 4.0 8 0.8 3. 9 0.6 Second level.8 0.5 0.7 1.7 0.7 3.3 1.0 3.8 1. 4.3 1.1 4.5 1.3 4.7 1.1 4.1 1. 4. 1.0.4 0.7 1.3 1.9 0.5.6 0.8 1.8 0.5 1.0. 0.6.7 1.0 4.1 1. 4.1 1.3 4.5 1.1 4. 1.3 4.6 1. 3. 1.0.9 0.6 0.9 1.4 0.5.6 0.6 11 1 Third level 13.0 14 0.9.7 Fourth 15 level 1.7 0.8.0 1 3 10 R P N M L K J H G F E D C B A 1.8 0.9 1.5 0.8 1.6 0.7. 0.5 1.3 0.5. 0.5 1.4 0.5.0 0.5 1.3 0.7.0 0.8 1.9 0.9 1.7 0.9 4.0 0.9.5 0.7 1.0 0.9 0.7 1.1 0.9 0.9 0.6.1 0.7 3.3 0.9 1.7 0.9 3.5 0.8.7 0.5 0.6 0.9.5 0.7 3.0 1.1 4.0 1. 3.9 1. 4.4 1. 3.4 1.1 3.3 0.7 1.3 1.0.1 0.5.9 0.8.5 0.7.1 0.7 1.4 0.6 1.0 0.5.8 0.7.8 1.0 3.5 1.0 3.0 1.0 3.1 0.7 1.3 0.5 0.9 0.9 1.5 0.7. 0.7 1.8 0.8.7 0.7 1.3 0.5 0.5 0.8.3 0.6 1.9 0.7. 0.6 1. 0.8 1. 0.5.3 0.7 1.6 0.8 1.7 0.9 4.4 0.9.6 0.7 1.4 0.5 0.8 0.6 0.6 0.8 1.0.3 0.7 3.8 0.9 1.6 0.9 Fifth level 1 3 4 5 6 7 8 9 10 11 1 13 14 15 1.7 0.7.8 0.8.7 0.6.9 0.8.5 0.6 3.1 0.8 1.7 0.7 0.7.6 0.5 1.6 0.5.5 0.5 1. 0.5.3 0.5 1.6 0.7 1.6 0.8 1.4 0.9 0.7 3.5 0.8 3.0 0.6 3.6 0.8.8 0.6 3.0 0.8 1.8 0.7 With 1,000,000 neutron histories R P N M L K J H G F E D C B A 0.8 1.7 1.9.1 1.7.1 0.9 0.9 0.7 0.5 1.4 0.6 1.4 0.7 1. 0.6 0.7 0.9 0.8.7 1.4 0.7 0.7 0.5 1.1.0 0.8 1. 1.8 0.6 1.8 1.3.0 1.0 0.8 1.0 0.7 0.9 1.6 1.0 1.1 1.9 1.8.5 1.9.3 1.1 0.9 0.9 1.0.8 1.9 0.5 1.7 1.7.7.3.9.3.3 1. 1.0 0.8 1..5 0.9 0.6 1.1 1.5.5..9.6.8.1.1 0.6 1.1 1. 3..1 0.8 1.8.0.5.5 3.0.4.7 1.5 0.9 1. 1.8.5 1.1 0.6 1. 1.4.4..8.3.8 1.9.0 0.5 0.7 1.8.4 1.8 1.4 1.6..1.6.0.1 0.8 0.6 1.3 1.9 1.6 1.0 1.1 0.5 1.6 1.5 1.9 1.7 1.9 0.7 0.5 0.6 0.5 1.3 1.0 1.7 0.5 1.4 1.0 1.3 0.6 1. 0.5 0.9 3.0 1.6 1.1 0.7 1.3.4 0.7 1.1 0.9 0.9 1.8 0.9 1.7 1.5 0.6 0.8 0.5 1.8.5.3.8.0 1.9 0.9 With 10,000,000 neutron histories 10 8 6 4 0 4 6 8 10 10 8 6 4 0 4 6 8 10 1 3 4 5 6 7 8 9 10 11 1 13 14 15 1 3 4 5 6 7 8 9 10 11 1 13 14 15 R P N M L K J H G F E D C B A 0.9.0.0.4 1.8.3 1.0 1.5 0.6 1.1 0.5 0.6 1.5 0.5 1.6 0.6 1.3 0.7 1.3 0.5 1.8 0.6 1.4 0.6 3.1 0.5 1.9 1.0 0.5 0.5 0.8 0.6 0.6 0.6 1.6.4 0.5 1.6 0.6 1.7 0.5.3 0.6 0.5 0.7.1 1.7.3 1.1 1.0 1.1 1. 1.5 0.5 1.8 1. 1.4 0.6..4 3.1.5.6 1. 1.1 0.5 1.1 1.1 3.1.0 0.7 0.5.0. 3.3 0.5.9 0.5 3.5 0.5.8.6 1.3 1.1 0.9 1.4.6 1.0 0.8 1.4.1 3.1 0.5.8 0.5 3.3 0.5 3. 0.5 3.3 0.5.6.4 0.7 1.0 1.3 3.5. 0.6 1..4.6 0.5.9 0.5 3.0 0.5 3.4 0.5 3.0 0.5 3. 1.9 1.1 1.3.1.5 1. 0.8 1.5 1.9 3.0 0.5.7 0.5 3.1 0.5.8 0.5 3.4 0.5.4.3 0.6 0.7 1.9.7.0 0.5 1.7.1.8 0.5.7 0.5 3. 0.5.5.4 0.9 0.7 1.4.1 1.7 1. 1.4 0.5 1.9.0.5.. 0.8 0.7 0.9 0.6 1.4 1.4 0.5. 0.7 1.7 1.4 1.6 0.7 0.7 0.6 1.7 1.0 0.5 1.5 0.6 3.5 0.5.1 1.4 0.6 0.5 1.0 1.8.9 0.5 1.3 0.6 1.7 0.6 1.3 0.5 1.1 1.9 1.0 1.9 0.5 1.6 0.7 1. 0.5 1.3 0.6 1.9.7.3 3.1.1. 1.0 With 5,000,000 neutron histories R P N M L K J H G F E D C B A 0.6 1.4 1.8 1.7 1.6 1.7 0.7 0.8 1.3 0.5 1. 0.5 1.1 0.7 0.9 0.6.4 1. 0.7 0.0 0.6 0.5 0.0 0.9 1.7 0.7 1.0 1.6 0.6 1.7 1.3 1.9 1.0 0.8 0.9 0.5 0.9 1.5 0.8 1.0 1.8 1.8.4 1.8. 1.0 0.9 0.6.5 1.8 0.0 1.6 1.6.6..8.. 1. 1.0 0.0 0.7 0.8.4 0.7 0.5 1.0 1.4.4..9.6.6.0.0 0.5 0.9 1.1.8 1.9 0.8 1.7 1.9.5.5.9.3.5 1.4 0.9 1.0 1.4.3 0.9 0.5 1.1 1.3..1.7..7 1.8 1.9 0.6 1.7.1 1.8 0.0 1. 1.5.1.0.5 1.9.0 0.7 0.6 0.0 1. 1.5 1.5 0.8 1.1 1.5 1.4 1.8 1.6 1.8 0.7 0.5 0.5 1.1 0.7 1.4 1.3 1.0 1. 0.6 1.0 0.7.8 1.4 1.1 0.0 0.0 0.6 1.1. 0.5 0.9 0.7 0.7 1.7 0.7 1.6 1.4 0.6 0.5 1.6.1.1.4 1.8 1.6 0.7 With 50,000,000 neutron histories 10 8 6 4 0 4 6 8 10 10 8 6 4 0 4 6 8 10 Figure : Differences in radial assembly power distributions at core mid-plane between Serpent-ARES and Serpent 3D reference solution. Group constant generation with different number of neutron histories. Top value: maximum difference, bottom value: standard deviation of differences. Calculations performed with 0 sets of independently generated group constants.
Second level Estimation of computational cost The overall calculation time was extrapolated from single assembly calculations: Running times from Serpent burnup calculation 9 assembly types, 9 history calculations, 16 burnup points 7 or 54 Third branch level points, depending on assembly type (all combinations from state-point Fourth matrix) level Fifth level Table 3: Running times of single history and single branch for the representative fuel assembly (.4 wt-% 35 U fuel with 1 burnable absorber pins), together with the estimated total running time required for producing the full set of group constants. All calculations were run with 1 CPU cores (3.47 GHz Intel Xeon workstation). Neutron Running times (min) Estimated total (days) histories T burn T bra 1M 30.5 0.5 18.5 5M 91.5 1.8 6.8 10M 169.5 3.5 119.3 50M 790.9 16.6 566.9
Observations and conclusion of the study: Second level Estimation of computational cost Increasing the number of neutron histories beyond 10 million in homogenization does not significantly change the results of the ARES calculation Random Third asymmetry level in ADF s can inflict large systematic tilt in the core power distribution, Fourth which level can be avoided by averaging over the symmetries 15 Covering all state Fifth points level was estimated to take a few CPU months of calculation time, which can be significantly reduced by running separate history calculations in different computer nodes In reality, the overall computational cost is considerably lower, since all state-point combinations are not used 15 Serpent has an option for this in the ADF card
Second level Using continuous-energy Monte Carlo codes for the task has several advantages over traditional deterministic approach Third level Serpent was Fourth originally level designed specifically for for spatial homogenization, which still remains as one of the two main focus areas for future development Fifth level Summary and conclusions Traditional calculation sequence used for modeling operating nuclear reactors is based on spatial homogenization Implementation of an automated sequence for performing branch calculations is under way The results have been promising, but more user experience is needed! Serpent seems like a viable option for spatial homogenization, even when burnup and all state points are covered, but a full-scale demonstration is yet to be accomplished
TECHNOLOGY FOR BUSINESS