The ANSWERS Seminar, 22nd-24th May 2007

The ANSWERS Seminar was held at the Durley Hall Hotel in Bournemouth, a location near to the superb Dorset coast and in close proximity to the town centre.

Over ninety 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 emphasis of the first day of the 2007 ANSWERS Seminar was Radiation Shielding. Radiation Shielding applications are primarily carried out using the MCBEND and RANKERN radiation transport codes. A variety of application and developmental presentations were performed by both Serco Assurance personnel and guest presenters.

The morning started with a presentation on MCBEND 10A Release Update 1 provided by Geoff Dobson. This gave an overview into errors found in the original release of MCBEND 10A, how these errors have been rectified and described how testing had been carried out to check the corrections made. Particular emphasis was made on changes between MCBEND 10A and MCBEND 10A_RU1 that will impact on the user. The presentation was followed by an applications talk by Roger de Wouters of Tractabel.

Roger provided some insight into how PANTHER and MCBEND are being utilised to calculate source level ex-core detector responses. The next presentation was carried out by George Wright, who spoke about generation and testing of IRDF2002 dosimetry and covariance data libraries. These libraries will be available with MCBEND 11.

Following the morning break, a presentation given by David Picton described how the Diffusion Solution to the Boltzmann Transport equation may be used in MCBEND to provide a deterministic approach to radiation transport. This was followed by a presentation made by Dennis Allen on MCBEND flux calculations that had been carried out for Oldbury. Dennis demonstrated that new features in MCBEND 10A significantly cut down on the number of lines of input required for the whole reactor model, compared to previous work. Adam Bird and Tim Fry ended the morning session with a demonstration of the features available in Visual Workshop. Visual Workshop is expected to provide a single replacement for all various graphics packages available from Serco Assurance. A beta version of Visual Workshop was made available on CD to Serco Assurance customers for a trial period.

The afternoon session started with a Grid Launchpad GUI demonstration and description of MCBEND parallelisation, provided by Adam Bird. The demonstration was followed by an application talk provided by Lewis MacFarlane from Sellafield Ltd entitled Long Term Advanced Gas-Cooled Reactor Fuel Storage in Fuel Storage Ponds. The next presentation was carried out by Adam Bird describing the CAD geometry interface, which should enable CAD geometry compatibility in ANSWERS codes.

Following the afternoon break, a presentation given by Victoria Hooley described enhancements made and testing carried out on the new scoring unit (Unified Tally module), which will be made available in MCBEND 11. This was followed by a further presentation made by Geoff Dobson describing and demonstrating the use of ANSWERS customer web pages for shielding codes. Finally, a presentation made by Pat Cowan described the current status and future plans intended for ANSWERS codes, in particular MCBEND, RANKERN and Visual Workshop.

The second day of the meeting, the Reactor Physics day, started with a presentation on recent work to extend the input and output features of WIMS. Parameter variables are now available in the input language as well as equations and an additional level of nested cycles. Additional graphics facilities include a LaunchPad for running WIMS, graphics output options from the EDIT module and visualisation of the new FG CACTUS3D geometry definition through VisualWorkshop. The second presentation was given by Kevin Hesketh of Nexia Solutions who described the use of WIMS to model the core and fuel cycle of a 5 batch PBMR.

This work is being performed as a contribution to the FP6 PUMA (PlUtonium and Minor Actinide) project. The current aim is to evaluate detailed inventory data for fuel cycle scenario benchmarks, however, the studies have demonstrated the potential of PBMRs for actinide burning although there are difficulties in maintaining reactivity with Pu or Pu/MA fuel. The final presentation before the mid-morning coffee break detailed advances with the MAX module. The MAX method applies direct Monte Carlo scoring to perturbations such that only the perturbation is scored rather than taking the perturbation as the difference between two Monte Carlo calculations. This method is fast and the stochastic uncertainties only apply to the perturbation itself. New advances to speed convergence of the method include the use of projection operators.

The first presentation following the coffee break was given by Paul Bryce of British Energy and described the use of WIMS as a tool for failed fuel prediction. The primary coolant gas in a PWR is monitored by gamma spectroscopy for particular radioisotopes to identify any possibility of fuel failures. If detected each fuel element is sniffed when offloaded from the core and any failed fuel discarded. During operation, however, some primary coolant analysis is performed in order to identify the most probable set of assemblies containing the failure using the ratio of the Cs137 to Cs134 activities. A legacy method failed to identify the correct group of assemblies containing the failure in Cycle 8 and an independent review of the method, using WIMS9, has been performed. The review showed the WIMS9 method to be much better and this method although needing further validation shows promise for application in the future. The next presentation was given by Pavel Mikolas of Skoda and considered the comparison of measured and calculated isotopic compositions from eight VVER 440 3.6 w% enriched UO2 samples irradiated in the Novovoronezh IV unit between 1987 and 1991. Overall the agreement between measurement and calculation was poor and better agreement had been obtained for other irradiations. It is considered that the problem is not with the analysis codes or data but is mostly due to problems with the measurements themselves.

The final presentation of the morning outlined advances in nuclear data measurements and analysis. In particular the PhD programme of Tim Ware (Serco) was outlined which consists of performing cross-section measurements at Geel in Belgium using time of flight and capture/transmission yield measurements. The measurements will be analysed using methods developed by Moxon to yield the cross-section resonance structure. It was emphasised that the UK must maintain this type of measurement and analysis expertise as nuclear data has a fundamental importance in many aspects of the nuclear industry.

The first two presentations after lunch were given respectively by Christophe Schneidesch and Roger de Wouters of Tractebel Engineering SUEZ and described the development of the WIMS supported Panther core monitoring and reactor analysis package CORMORAN (CORe Monitoring and Reactor Analysis) for the Belgium nuclear power plants. Christophe presented an overview of the permanent core follow / core status display package, the core strategy and forecast package and the in-core flux map analysis package. The presentation also described the secure IT network and backup requirements needed to avoid data loss and protect data integrity. The system is being validated by application to the operating history of the Belgium reactors as well as against operational transients. Analysis of results for the evolution of the core axial offset show good agreement for the Tihange 1 reactor (12ft high core) but some oscillatory behaviour for the Tihange 3 reactor (14ft high core). However, taller cores are more prone to axial sensitivity. Currently it is expected to re-normalise the core follow calculations to the measured conditions every 48 hours. The analysis of operational transients again shows good agreement for the 12ft high cores but presents challenges for the taller cores.

Roger presented the analysis and development of power reconstruction and flux map measurements. Flux maps are taken monthly on each of the reactors with reaction rates being measured using miniature fission chambers. These are then processed to reconstruct the 3D power distribution and to check key parameters such as the axial and assembly hot spots. The current processing code Rabbit is now out of date and is being replaced by Adjust developed within the frame of CORMORAN. Several improvements to the route were described including axial model refinement and explicit grid modelling as well as a method for the reconstruction of powers in non measured locations using coupling coefficients. It is concluded that the uncertainty on axial and assembly hot spots is 4% and 5% respectively for core states near 100% nominal power. The final presentation before the afternoon tea break described a new set of features in WIMS related to temperature calculations. An iterative calculation was described where initial guess temperatures were used to resonance self shield the cross-sections for a pin geometry and following a flux solution the new WIMS TEMP module was used to calculate the pin temperatures. The updated temperatures were then passed back to the next round of iteration using the new feature of storing temperatures on the WIMS interface. Note that this also means no temperature data need be given to the RES module as it is placed on the interface by PRES. The resonance self shielding calculation in PRES and RES has also been updated to allow the same resonance nuclide to be at different temperatures in different materials. This feature allows the treatment of radial temperature profiles in the pin. Finally the iterative calculation was terminated using a new interface compare option in the INTER module which will compare selected parts of two interfaces within a given tolerance, in this case temperatures, and optionally stop the calculation if they are the same. This compare feature is also useful for QA checking of validation calculations, failures against reference cases being independent of the print level used and being flagged by a failed interface compare message.

The first presentation in the final session of the day described the current state of the CACTUS3D characteristics method flux transport solution module. CACTUS3D is due for release with WIMS10 in December 2007 and has been fully tested against results from CACTUS2D. There are differences in the tracking algorithm between the 2D and 3D versions and these were described in the presentation. As well as the geometry description as given to CACTUS2D a new fractal geometry (FG) input, in common with the MONK and MCBEND Monte Carlo codes, is also allowed. This form of input has the advantage that the geometry can be viewed using the VisualWorkshop graphical package providing views across slices within the geometry as well as views by material type etc. The FG geometry is also much more powerful for describing complex geometries and allows more natural sub-meshing, sub-meshing within a cylinder will be in the form of annuli for example. The next presentation was an invited presentation by Les Hutton who retires from Serco in the Autumn. Les presented a talk on reactor physics over 4 decades. As well as describing his own history within the nuclear industry Les looked back at the major changes in the working environment. Some were less positive such as the loss of many of the experimental facilities but others such as the improvements in computer performance allowed greater complexity and precision for experimental and reactor analysis. Les finished his presentation by discussing some of the problems and challenges still faced by the industry.

The final physics presentation of the day was given by Robert Mills of Nexia Solutions and reported calculations estimating burnup at the end of PWR fuel rods using CACTUS3D. The TRAIL module was included in the calculational route to provide cross-sections for the FISPIN code to calculate detailed nuclide inventories. Plutonium burnup at the lower end of the pins was shown to be much larger than the assembly average burnup while burnup at the top of the pins was slightly less than the assembly average. As expected power production shifted toward the end of the pins as irradiation continued due to the burnout of fuel at the centre of the pin.

The day was completed by a short presentation on the contents of WIMS10, due for release at the end of 2007, relative to WIMS9 as well as describing future plans for the enhancement and development of the WIMS code.

The day was closed with an informal discussion and problem clinic session.

The third and final day of the meeting, the Criticality day comprised the Criticality related technical presentations. Once again there was good representation from different areas of the criticality community: analysts, assessors, regulators and code developers.

The technical presentations began with a description by Les Hutton of a new MAX development for MONK that uses Monte Carlo tracking to provide results from a perturbed system. The basic perturbation method was described, including a list of all the properties of the model that could be perturbed. This was followed by a presentation from Simon Connolly of Sellafield Ltd on the sensitivity of MONK to nuclear data parameters. This studied the effect on the results of three fast systems to changes in theses parameters the conclusion being that the number of angular bins and the subgroup relative width can affect the results and must be chosen wisely. The final talk of the first session was from Kim Coggins of Nexia Solutions and she described the safety assessment work carried out on a recently refurbished decontamination facility. This described the various equipment installed and processes that must be followed and how the safety case must look at each of these individually and their possible interaction to ensure plant and operator safety.

After the morning break the presentations continued with an AWE presentation by Nigel Tancock on historic experiments carried out at AWE and the difficulty of finding all the necessary data after many years. This was a detective story with a successful outcome, but the important point raised was the need to urgently try and find all the required data before it is shredded or the now aged experimentalists die off. This was followed by a talk from Paul Hulse of Sellafield Ltd on the CODEMORE tool for distributing MONK (or MCBEND/WIMS) looping calculations across a network of computers. This described the WORM tool and how CODEMORE manages the queues of jobs running the inputs created by WORM to ensure that machines are not flooded. This talk also gave a brief description of MONKCHECK a tool for checking many MONK outputs files for typical error reports and providing a succinct table for the user to review. The next talk was from Alwin Wiederhold of British Energy, and he described the burnup credit work they have done to support a safety case for the Sizewell B wet storage pond. This included studies of the bounding conditions and axial depletion profiles. The final presentation of the morning session was from Adam Bird of Serco Assurance and he gave a brief talk on the new features of Visual Workshop followed by a demonstration of these features. A disc with a beta version of Visual Workshop was made available for attendees to take away and try.

After lunch the programme continued with a description by Bernard Franklin of Serco Assurance of a benchmark study using the PROTEUS hole in MONK. This studied a series of cores comprising fuel and graphite pebbles (similar to PBMR) in different arrangements and compared the MONK results with experiment. The conclusions were that MONK gives good predictions for all the cores studied and predicts the rod worths to within 8%. This was followed by a talk from Huo Xiaodong of the Beijing Institute of Nuclear Engineering on the criticality analysis of the Qinshan Phase II NPP Fuel Pool. The talk gave an overview of the pool and fuel assembly stored therein.

Scenarios studied included a dropped fuel assembly that sits between other assemblies (several possible configurations) and the optimum moderated system when dry storage is normally used. The conclusions were that the dropped assemblies were all within acceptable criteria, and the peak keff for water ingress is just over 0.9 and occurs at 0.04g/cm3. The final talk of this session was again from Adam Bird, and this described the work done to import CAD models into MONK to remove the need to translate engineering CAD models into FG/HG input. The talk detailed the difficulty in automatically translating between these data types and looked at both the tetrahedral mesh and the NURBS surface methods for importing models. There exists a hole in MONK to handle the tetrahedral mesh and are studying the best way to handle NURBS surfaces. Adam noted the need for a good CAD model it appears that many models have mismatched surfaces which cause problems when tracking.

The final session on the final day of the 2007 ANSWERS seminar began with a talk by Jim Gulliford of Nexia Solutions on a risk-informed approach to criticality assessment. This talk discussed a range of limits as part of the ALARP framework and how some of these cannot be taken in isolation the need to consider the consequences of lower package limits such as increased operator handling dose must be included. An example showed how an understanding of the system modelled instead of blindly using worst-case values for all the parameters could allow a package limit to be raised from 2kg to 18kg thus reducing transportation and operator risk. The session closed with a summary of the current status and future plans for MONK, including details of the MONK9A_RU1 code release to fix a number of errors including the HEXARRAY bug, and once again a request from ANSWERS to users for development ideas that we can put into MONK to aid them with their work.


Thank you for letting us use cookies to improve your experience of this website. Please read our cookie statement for information on how we use cookies and how to restrict their use.