The 2009 ANSWERS Seminar was held at the Wessex Hotel in Bournemouth, a location near to the superb Dorset coast and in close proximity to the town centre.
Over one hundred delegates attended all or part of the three-day meeting, with the programme covering the shielding, reactor physics and criticality technical areas. Programme topics included ANSWERS code developments, code applications, related general interest issues and software demonstrations. In addition to the seminar itself, parallel meetings of the UK Shielding Forum, the UK Working Party on Criticality and the UK Nuclear Science Forum were hosted at the same venue.
The first day of the 2009 ANSWERS Seminar was Radiation Shielding, which was hosted by Pat Cowan, ANSWERS Applications Area Manager for Radiation Shielding. Pat introduced a variety of presentations on the recent developments in the ANSWERS Shielding codes and supporting tools and a number of applications using the ANSWERS codes.
The morning started with an applications presentation on “The use of Novel Features in MCBEND to predict AGR Core Restraint Dose” by Adrian Huggon (Serco, Quedgeley). The talk concentrated on the geometry features in MCBEND – Window Parts and Hole Geometries – combined with neutron source data from PANTHER core follow calculations, that have enabled the development of a highly detailed reactor model. This presentation was followed by a talk from Simon Warburton (NII) on “Legal Limits, Design Targets, BSLs and BSOs”. This summarised the legal limits for radiation doses to workers, trainees and members of the public. However the main point of the talk was to emphasise it is the responsibility of the licensee to show that radiation doses must be ALARP.
The next presentation was given by Adam Bird on “Testing of the Tetrahedral Mesh Import Capability for importing converted CAD files into MCBEND“. This was an update on last year’s presentation and the investigations that have been carried out to determine the causes of small discrepancies in results between MCBEND calculations run with traditional FG and the Tetrahedral Mesh. Results from a ‘hybrid’ model with both FG and the tetrahedral mesh were also shown.
Following the morning break, Geoff Dobson gave a presentation on a development he has been working on in MCBEND – “BINGO Adjoint Data”. Geoff started with some background on the BINGO collision processor that will be in MCBEND 11. Then he presented the theory behind the adjoint method, and what needs to be done to generate adjoint data. This was followed by an interesting presentation given by Dean Thornton (Serco, Quedgeley) entitled “A Case Study addressing Unexpected Results”.
Dean presented some unusual results he had recently seen in fast neutron dose survey on a core barrel near horizontal former plates. He came up with a possible explanation for the results, but invited comments from the audience. George Wright ended the morning session with an applications talk entitled “Shielding Design of the NPL Clinical Linac Facility“. George described the new facility and the Monte Carlo modelling work he carried out to demonstrate that the final shielding design met with the dose-rate criteria specified by the customer. Some iteration in shield design was required in order to meet the criteria.
The afternoon session started with a demonstration of the latest features available in Visual Workshop given by Adam Bird. This included direct screen measurement, displaying Forced Flight interfaces, Complex Source and Scatter/Reflect bodies as well as many other new features. This was followed by an applications presentation given by David Morgan (Serco Quedgeley) entitled “A New Approach to Graphite Dosimetry”. David described a detailed reactor model that has been set up using a number of new features in MCBEND 10. Next Geoff Dobson described a new capability that is being developed in MCBEND during his talk “Incorporating (gamma,n) reactions in the MCBEND All-in-One coupled capability”. This is still in an early stage and has currently only been implemented for a single nuclide, but the route has been demonstrated.
Following the afternoon break, a presentation given by David Picton described “Activation Calculations with MCBEND”. This talk was initiated by a number of helpline queries on the subject. David concentrated on the calculation of activation products in steel and mentioned a couple of different methods for doing such calculations.
Next a presentation entitled “Recent Experiences with Leakage Files on Large MCBEND models” was given by Anton Murfin of Sellafield Ltd. This described work using MCBEND Leakage files which considerably cut down the run times for long calculations, and how the ‘TIMES’ option was initially misunderstood by users. Finally, Pat Cowan ended the day by describing the current status and future plans for the ANSWERS Shielding Area.
The second day of the 2009 ANSWERS Seminar, the Reactor Physics day, was hosted by Tim Newton, ANSWERS Technical Director.
The day started with a joint presentation by Nigel Davies and Simon Richards, both of Serco. The presentation covered new developments of MONK in WIMS. Nigel discussed MONK code modifications that enable material- and incident-group-dependent fission spectra to now be used in calculations. This capability removes previous modelling approximations related to the treatment of fission spectra in MONK and can thus improve the accuracy of reactivity predictions, particularly for mixed-fuel systems. Simon presented the new Unified Tally (UT) scheme and a treatment of Shannon entropy in MONK. Existing MONK methods for scoring quantities such as neutron flux can be restrictive, leading to constraints on how model geometries are defined. The UT scheme allows general scoring bodies to be overlaid onto a MONK geometry model, irrespective of how the model has been specified. An example of the application of UT to scoring individual pin fluxes in a PWR whole-core model was given, thus demonstrating the powerfulness and flexibility of the method. Shannon entropy was proposed as a new metric to determine the degree of convergence of a flux distribution. Estimates of the Shannon entropy of a system can also indicate how significantly the flux distribution varies with spatial position.
The next presentation was given by Pavel Mikolas from Skoda. This covered pin-power analyses for the Temelin, Unit 1 core (type VVER-1000). Cross-section data were prepared using WIMS9; core calculations were made using the MOBY-DICK finite-difference diffusion solver. The deterministic results were compared with results from reference MCNP calculations. Good agreement was found between the WIMS9/MOBY-DICK and MCNP pin-power estimates. Improvements in the WIMS9/MOBY-DICK results could be made by applying a correction factor to the diffusion coefficients to account for mesh irregularities. The final presentation before the mid-morning coffee break was given by Paul Bryce of British Energy. Paul discussed 2D whole-core PWR calculations using CACTUS. Sensitivity analyses for tracking parameters and meshing were discussed. It was concluded that CACTUS is capable of perform this type of large-scale calculation, although further developments would be needed to improve its usability in this area. Suggested areas for improvement included new acceleration methods and moving to 64-bit code to overcome data handling issues.
The first presentation following the coffee break was given by Matthew Worsley of AMEC. This presentation covered a route for performing calculations of pebble-bed reactors using WIMS9. The neutronics calculations had been coupled to an AMEC-written thermal hydraulics code, thus allowing temperature feedback effects to be treated. Roger Thetford of Serco gave the next presentation, discussing the TRAFIC code. TRAFIC contains detailed mechanistic models of physical processes and properties and can be used to model fuel behaviour under operational and accident conditions. Roger highlighted some previous applications of TRAFIC, including transient analyses for PFR, Cabri and VIPER and fuel design calculations for CAPRA/CADRA, CONFIRM and AFTRA. The next presentation was given by Brian Turland of Serco. Brian presented new ways to use WIMS for calculating effective reflector data for CANDU and Magnox-type reactors. The reflector materials in these reactors typically have a low absorption cross-section. It was found that the existing “albedo” method for deriving LWR reflector data was inadequate for modelling these systems; after correction the reflector cross-sections were often found to be negative. The new approach uses neutron currents derived from a detailed transport solution for a model of the core/reflector interface. A new fitting process, based on converting β-type albedos to λ-type albedos, ensures that no negative reflector cross-sections are produced. Reflector data derived using the new methods were used in a diffusion code; the results from which were found to be in good agreement with reference Monte Carlo solutions.
The final presentation of the morning was given by Paul Bryce and discussed modelling of sub-critical AGR and PWR cores with WIMS9/PANTHER. The only neutron sources considered in the calculations were spontaneous fission, sources from Be/Sb and (α,n) reactions in the oxygen of UO2. Data for calculating the sources were derived from WIMS and JEFF. It was found that the measured sub-critical flux/power levels could be reproduced to within a factor of two. Improved agreement with measured flux levels could be achieved by adjusting the magnitude of the sources. Paul proposed that there may be a deficiency in the approximations used for calculating the sources. A member of the audience suggested that the discrepancy could have arisen as a consequence of uncertainties on the nuclear data used to calculate the sources. For example, uncertainties on some curium data could be as large as 50%.
After lunch, the first presentation was given by Glynn Hosking of Serco. This covered the generation of FISPIN data libraries by using WIMS. Typical usage of the WIMS/TRAIL cross-section library preparation route was discussed. Differences between versions of the TRAIL code, and their compatibility with different versions of WIMS, were highlighted. Next, Christopher Dean of Serco presented an overview of the data libraries that will be issued with WIMS10. Three WIMS libraries will be made available: i) a JEF2.2 library as issued with WIMS9 but with added intrinsic source data ii) a JEF2.2 library with corrected Pu239 data, intrinsic source data and deep-burn data iii) a JEFF3.1 library featuring extended data for burnable poisons, gamma source data, heating data and extended and standardised temperature tabulations. JEF2.2 ECCO libraries, for use with the WIMSECCO module in WIMS10, will also be made available. The JEF2.2 ECCO libraries are consistent with the JEF2.2 WIMS libraries. Versions of the TRAIL/WEDIT libraries of infinite-dilute cross-sections will also be made available. The final presentation before the afternoon tea break was based around a demonstration of the Panache code, given by Tim Fry of Serco. Panache is an interactive user interface for PANTHER, originally developed by British Energy. It is designed to facilitate fuel cycle design calculations for AGR and PWR systems. The current version of Panache is only compatible with SunOS due to the use of the Sun OpenLook toolkit for the user interface. Panache has now been re-engineered to use Java as its toolkit, meaning that it can be ported to Linux and new versions of Solaris. Tim demonstrated various features of the code, including user-controllable features for displaying neutronics parameters overlaid onto an illustration of the core map.
The first presentation in the final session of the day was given by Dave Powney of Serco. Dave gave an overview of the WIMSBUILDER code. Typical examples of WIMSBUILDER input files were presented, alongside code features such as macro substitution and conditional input. Dave presented the available versions of WIMSBUILDER as being the AGR, VVER and single-cell and supercell PWR variants. Prototype versions of WIMSBUILDER are available for other types of cluster geometry, namely RBMK and CANDU systems.
Sheldon Hall of Imperial College next presented details of his EngD project to study the development of nodal methods for the analysis of PWR cores with advanced fuels. This project is joint-funded by British Energy and the EPSRC. An overall aim of the project is to develop PANTHER so that it can accurately model LWR cores containing mixtures of MOX and UOX assemblies. Sheldon highlighted the current problems with the PANTHER treatment of MOX as being: i) transport effects – higher-order approximations than diffusion theory are needed to accurately treat the interface between UOX and MOX assemblies ii) few-group effects – two-group data for MOX fuel may contain significant condensation errors, it is possible that more energy groups are required. The planned EngD programme will consist of several phases. The first phase will be for Sheldon to study SPN finite-difference schemes and the implementation of nodal methods. Another phase will concentrate on developing production methods for PANTHER. Various angular discretisation methods (SN, PN and SPN) and associated matrix solver methods will be investigated. The aim is to determine the fastest embedded solver for the required accuracy. Sheldon will also perform research for the Imperial College group, specifically in the areas of proper orthogonal decomposition and linking coarse, core-scale solutions to detailed, assembly-scale solutions. Initial work by Sheldon has considered the use of MATLAB to investigate differences between diffusion theory and SPN for the TWIGL benchmark.
The day was completed with a short presentation from Tim Newton, the host of the Seminar reactor physics day. Tim stated that, unless a customer has specific QA requirements, WIMS10 will be released as a beta version so that it can remain open for development. Despite this, the beta release will have full testing associated with a QA release. A summary of current WIMS developments was given as:
Tim also provided a re-cap of the new features in WIMS10:
The final day of the 2009 ANSWERS Seminar focused on Criticality and was hosted by Malcolm Armishaw, ANSWERS Applications Area Manager for Criticality. The day began with introduction by the ANSWERS Manager Paul Smith and a safety talk from Karen representing the hotel. The technical programme commenced with a MONK Status presentation from Malcolm Armishaw giving an update on the code, any memos since the previous seminar and an overview of some of the current code development tasks. Malcolm also gave the next talk which was an overview of the MONK BINGO Validation work, describing the results obtained, and that a report is now available. The conclusion from the report is that the BINGO library available with MONK version 9A is now considered to be at production status level and no longer just for evaluation purposes.
This was followed by a talk from Lauren McDonald of NNL on the use of MONK8B in the refurbishment of their Nitric Acid Wash Facility, in particular looking at the geometries of their washing machines and liquor storage tanks. The final talk of the first session was from Keith Searson of Sellafield Ltd who described the developments to the CAD import into MONK and MCBEND. Keith described the two main supported representations, and then detailed the relative speeds of the various methods available, noting the trade off between precision of the model and speed of tracking though it.
The first presentation of the second session was on the prototype of a version of MONK with UT scoring and Shannon Entropy tallies. This was written by Simon Richards and presented by Nigel Davies. A detailed description of the implementation of the Unified Tally module in MONK was given, along with a series of test cases. This was followed by the use of the UT capability to score data for the Shannon Entropy to give a quantitative description of how converged the flux was in a calculation. This was followed by a talk from Malcolm Armishaw on the Verification and Validation of MONK, and covered both the existing process and how this will change with the availability of specific test hardware and Auto Testing software to drive it. The final presentation of the morning was from Jim Gulliford who described verification work for LEU systems aimed at establishing a simple test set that was not clouded by experimental uncertainties, an overview the NNL GEMSTONE 240 cpu computing resource, collaboration on the ARH600 project to regenerate/recast old data using new methods and data, closing with a discussion on the proposed WPC subgroup on codes to coordinate UK activities in feeding back data to code developers and assist in code testing.
The afternoon session began with a presentation by Derek Putley from Serco who described his ICASPA algorithm for improving critical and safe parameters by using buckling theory in an iterative scheme to derive the required parameters, and how this was accomplished using a DOS scripts and a combination of parameters and embedded files in MONK. The second talk was given by Tim Fry of Serco and was an update on the Visual Workshop visualization code and the improvements that had been made to it.
The final talk of the session was from Gary Palmer from Sellafield Ltd and was a review of the NUMDEN algorithms for Nitrate aqueous solutions. Gary described the proprietary code NUMDEN and its methods for determining solution compositions, then compared this with a range of other codes that also calculate solution density. The conclusion was that there may be up to 2-3% variation in solution density, but that this typically results in a small variation in k-infinity of 0.2%.;that there should be a benchmark algorithm validated by experiment; and there were other solutions for which uncertainty exists.
The first talk of the final session of the 2009 ANSWERS Seminar was given by Adam Bird in which he described his experiences of using the TETMESH hole for modeling imported CAD models. He highlighted some problems of using CAD such as there being a range of file formats to be handled, the process of translating a CAD model into a tetrahedral mesh, and how the code would track though the resulting mesh. Adam described some of the testing that had been done, noting that poor mesh selection can lead to significant changes to the results, in this case in an annulus where the mesh did not fit the material sufficiently well and for a series of source rods where a coarse mesh had a significant effect on the modeling of the rods. The next talk was from Nigel Davies of Serco who described the development of an option in the WIMS route of MONK to provide a material dependent fission spectrum, so the energy of the emergent sample is dependent on both the incoming energy and the material. This should give a much improved behaviour with Plutonium and MOX systems.
The final talk of the seminar was a description of the future development and testing plans for MONK. These included further BINGO work, nuclear data library developments (run-time Doppler broadening), auto testing, MONK validation, CAD developments, and an investigation of alternative parallelization schemes for MONK.
Malcolm Armishaw concluded the session and seminar, by thanking all the presenters for their efforts, all the attendees for their support, and hoped they had enjoyed this year’s seminar.
Customer feedback is very important to ANSWERS, and to this end we hand out critiques to all delegates, where various aspects of the seminar can be marked and specific points can be raised.
The number of returned critiques this year was much higher than from previous years, possibly due to a re-formatting of the questionnaire, making it more user friendly, or possibly due to the promise of a prize draw for all returned critiques! The winner of this year's draw was a delegate from NRG.
While feedback from the critiques was largely positive, we have taken on board the comments made, and are addressing suggestions for improvement for next year’s event. For example, better use needs to be made of the microphone and sound system. Also, the lay-out of the main room is likely to be modified, and before the next seminar, the room itself is to undergo major refurbishment.
ANSWERS would like to say a special "Thank you" to all those who returned critiques this year. We really appreciate your feedback and by taking on board your comments, we will be making improvements which will benefit future Seminars.
It is with great pride that next year we will be celebrating the 25th Anniversary of the ANSWERS Seminar.