Calculations to Support JET Neutron Yield Calibration:
Contributions to the External Neutron Monitor Responses Luka Snoj1, Brian Syme2, Sergey Popovichev2, Igor Lengar1, Sean Conroy3 and
JET EFDA Contributors*
JET-EFDA, Culham Science Centre, OX14 3DB, Abingdon, United Kingdom
1EURATOM-MHEST Association, Reactor Physics Division, Jožef Stefan Institute, Jamova cesta 39, SI-1000 Ljubljana, Slovenija, Luka.Snoj@ijs.si
2 EURATOM-CCFE Fusion Association, Culham Science Centre, Abingdon, OXON, OX14 3DB, United Kingdom
3 EURATOM-VR Association, Department of Physics and Astronomy, Uppsala University, Box 516, SE-75120 Uppsala, Sweden
ABSTRACT
Neutron yield measurements are the basis for the determination of the absolute fusion reaction rate and the operational monitoring with respect to the neutron budget during any campaign for the Joint European Torus (JET). After the 2010 changes of the JET plasma- facing materials (Carbon wall to ITER-Like Wall transition), confirmation of the neutron yield calibration will be ensured by direct measurements using a calibrated 252Cf neutron source deployed inside the JET vacuum vessel. In order to thoroughly understand the transport of neutrons from the vacuum vessel to the fission chamber detectors mounted outside the vessel on the transformer limbs and to computationally support the JET neutron calibrations project, we developed a simple but quick-running computational model of the JET tokamak for performing Monte Carlo neutron transport calculations.
Only 5- 10 % of the neutrons hitting the fission chambers penetrate the tokamak wall, other come via ports. The highest contribution to a certain fission chambers is via the closest port and the second highest contribution is via the port closet to the neutron source. If the port is blocked by a massive object, the fission chamber response is decreased by approximately the contribution of that port. It was observed that the torus hall wall significantly affects the external fission chambers response due to back scattering of neutrons.
The whole process of understanding and improving the knowledge of the neutron yield calibration for JET is of great interest for ITER, where the methods and procedures for calibrating the neutron yield monitors are still being developed, but the requirement is for 10 % accuracy in the fusion yield determination, as it is in JET.
1 INTRODUCTION
The Joint European torus (JET) is presently the world's largest magnetic confinement nuclear fusion research facility. It plays an important role in designing, testing and manufacturing of materials and components for the future fusion devices, e.g. the International Thermonuclear Experimental Reactor (ITER). At the end of 2009 JET entered a
* See the Appendix of F. Romanelli et al., Proceedings of the 22nd IAEA Fusion Energy Conference 2008, Geneva, Switzerland
shutdown state in which, through 2010, the Carbon plasma-facing wall will be replaced by an ITER-like wall (ILW) made of Beryllium, Tungsten and Carbon.
The replacement of wall will significantly affect the neutron yield measurements which are the basis for the determination of the absolute fusion reaction rate and the operational monitoring with respect to the neutron budget during any campaign.
After the wall transition in 2010, confirmation of the neutron yield calibration will be ensured by direct measurements using a calibrated 252Cf neutron source deployed inside the JET vacuum vessel. This calibration will allow direct confirmation of the external fission chambers calibration (which was the standard on JET originally) and provide the first direct calibration of the JET activation system inside the torus. The fission chamber (FC) system consists of sets of fission chambers mounted on three JET transformer limbs, which provide the time-dependent neutron yield used to assess a JET pulse. Activation detectors located inside the vacuum vessel are part of the JET activation neutron monitoring system, which determines the absolute neutron yields and hence the absolute calibration of time-resolved neutron yield monitors.
The last experimental calibration of the JET FCs was performed in 1989 [1]. Since then many JET systems have been added, some were removed and JET has been reconstructed several times. Hence the up-coming calibration is very important for accurate neutron yield measurements. The requirement for accuracy in the fusion yield determination is 10 %. In order to understand the important contributing factors of the calibration and to significantly improve the accuracy, a whole suite of calculations is required to support the JET neutron calibration project. Many are based on Monte Carlo modelling using advanced Monte Carlo transport codes, such as MCNP [3].
There are two existing MCNP models of the JET tokamak; homogenised, and detailed [2], both of which feature practically all neutronically important structures of the JET tokamak. Due to large amounts of detail in the actual JET fabrication, both models are simplified over the original engineering structure, but are still very complicated so that the calculations of neutron detector response take relatively long time to run (several days on a single core PC). In order to computationally support the JET neutron calibrations project, several hundred calculations are needed, which could not be feasible with existing models.
Hence we developed a simplified model of the JET tokamak featuring only the most important components.
The purpose of the study described in this paper is to perform a set of calculations using a simplified model of the JET tokamak to understand and estimate the various effects on the JET FC systems, such as: neutron source position, neutron source spectrum, importance of the ports, presence of large structures outside the tokamak (e.g. plasma diagnostics, plasma heating systems, iso-containers, etc.) and scattering back from the torus hall wall (room return effect).
The results of the calculations will serve as a guide to features which are important in the calibration. In addition the results will provide guidelines in optimizing the neutron source positions for detector calibrations. The results of the analysis are presented in the following sections.
2 CALCULATIONS 2.1 Computational tool
The calculations presented in the report were performed using the MCNP code (version 5.1.40), which is one of the most advanced and verified computer codes for Monte Carlo transport of neutrons and photons, [3]. All calculations were performed with ENDF/B-VI.8 [4] cross section library, which usually comes with the MCNP code package. The calculations
were separately performed using the FENDL 2.2 [5], JEFF 3.1 [6] and ENDF/B-VII.0 [7]
cross section libraries as well, but no statistically significant differences were observed. The
252Cf spontaneous fission neutron spectrum was taken from the IRDF-2002 [8]. The source intensity was 109 neutrons per second.
The error bars in all figures represent 1σ statistical standard uncertainty of the Monte Carlo calculations, unless stated otherwise.
2.2 Computational model
The structures (e.g. vacuum vessel, coils, mechanical structure, etc. ) of the simplified JET model are modelled as homogeneous layers. The cross section of the model in poloidal plane is rectangular in shape. However we conserved as much as reasonably achievable all the most important neutron transport parameters, i.e. component mass, material composition, major dimensions and vacuum vessel surface to volume ratio. The geometrical MCNP model of the JET tokamak is depicted in Figure 2-1. In the model the JET tokamak is surrounded by the torus hall wall.
Figure 2-1: Cross sectional view of the MCNP JET geometrical model. Horizontal section of the JET tokamak (right) and half of the vertical section of the tokamak (left).
3 FISSION CHAMBER RESPONSE 3.1 Source position
The FC chamber response was calculated in the FC detectors located on the transformer limbs. It is important to note that most of the results are presented as detector response versus detector position rather than detector response versus point neutron source position. The reason for this is mainly computational. If we wanted to calculate certain detector response versus source position, we had to run as many calculations as many source positions we would want to examine. As it takes approximately one day to run single calculation, such approach would be very time-consuming and inconvenient. Hence the approach of tallying several detectors at different positions in a single run is much more time-effective. Especially as the simplified JET tokamak model used in our calculations is rotationally symmetric around the vertical axis.
The first measurement in the neutron detector calibration campaign will be measurement of the FC response versus source position in the torus. As discussed above we actually calculated the FC response versus FC position. The calculations were performed for two limiting source positions, one in front of the E port (0 °) and one exactly between the E
and NE ports (22.5 °) (Figure 3-1 left). By performing additional runs with source at 5 °, 10 °, 15 °, 20 °, and 25° and considering the symmetry relations in the computational model of the tokamak we were able to reconstruct the FC response versus source position for the simple model without any perturbations (Figure 3-1 right).
The FC response versus detector position is symmetric as the detectors are arranged symmetrically in the computational model. The FC response versus source position is strongly asymmetric due to the position of the detector on one side of the transformer limb, which shields the detector from neutrons from ports behind that limb. A step like function of FC response versus position with steps around the ports suggests that the neutron source position relative to the ports is extremely important for external neutron detectors indicating that the majority of neutrons reach the detector via ports. This effect is thoroughly investigated in section 4.2.
Figure 3-1: Calculated FC response in transformer limb detectors versus detector position for point source at two positions (left). FC response versus source position (right).
3.2 Spectral effects
All calculations discussed above were performed by using a 252Cf spontaneous fission neutron source. As the DD and DT fusion neutrons on average appear at higher energies (2.45 and 14.1 MeV) we examined the effect of the neutron spectrum of the point source on the FC response. The next neutron calibration before the JET DT campaign will be most probably performed by using the DT neutron generator. Hence the results will be indicative for this calibration as well.
All azimuthal scans discussed in the paper were performed by using the open port, i.e.
with no closure plates covering the ports. The analyses of all effects have been performed by changing the port closure plate thickness as well. The results are altered but the main conclusions remain the same. Due to limited amount of space the results are not presented in the present paper.
The Cf-252 denotes the 252Cf spontaneous fission neutron spectrum from IRDF-2002 [8]. The 2.45 and 14.1 MeV denote isotropic mono energetic point neutron sources. The FC response versus detector position for three neutron spectra and two source positions is presented in Figure 3-2. When the neutron source is in front of the port, the differences in FC response due to different neutron spectra are relatively small and are on the order of 5-10 %.
It is interesting that the FC response to 14.1 MeV neutrons in the detectors closest to the source is significantly smaller than for the other two spectra. This can be explained by reduced backscattering from the torus hall wall, as these neutrons have higher leakage
probability. The FC response versus detector position changes significantly when the source is between the ports. In this case the FC response to 14.1 MeV neutrons in detectors closest to the port next to the source is significantly higher due to effectively thinner torus wall in the direction towards these detectors because of the port (see Figure 2-1) and due to higher torus wall penetration probability. Responses of other FCs do not differ significantly.
Figure 3-2: FC response versus detector position for three neutron spectra and two source positions; at 0 ° (left) and 22.5 ° (right).
4 IMPORTANCE OF TORUS STRUCTURES
In order to thoroughly understand the importance of various torus structures and where do neutrons come from into the FC detectors, we calculated the so called flagged FC response. Flagged FC response tallies only neutrons that have at least once in their lifetime crossed one or more flagged structures (cells). The purpose of the calculations was to examine:
• the room return effect, i.e. the neutrons that scatter back from the torus hall wall
• the importance of the ports relative to the position of the source and detector
4.1 Room return
The room return effect was examined by flagging the neutrons that crossed the torus hall wall boundary in their history, i.e. the portion of neutrons, detected by the FC, that are scattered back from the torus hall wall, versus detector position. Two major conclusions can be made. Firstly, the room return effect is stronger when the neutron source is in front of the port. This is mainly due to larger probability of escape from the vacuum vessel. Secondly, the share of back-scattered neutrons strongly depends on the position of the FC relative to the neutron source and ranges from 10 % for detectors closest to the source to 50 % for detectors on the opposite side of the torus. In other word 50 % of the neutrons, detected by the detectors on the opposite side of the torus from the source, have originated from other ports and have been scattered from the torus hall wall in their histories. Therefore any large items in the torus hall can potentially affect the FC response by blocking or redirecting the backscattered neutrons.
Figure 4-1: Percentage of the neutrons detected by the FC that scattered back from the torus hall wall, versus detector position. Two source positions: 0 ° and 22.5 °.
4.2 Port contributions
In the previous sections we observed that the neutron source position relative to the port is very important for the external monitor response. Therefore we examined that portion of neutrons contributing to individual detectors via various ports or the relative FC response via ports. We calculated the share of flagged FC response via a port relative to the total FC response (Figure 4-2). It can be observed that approximately 90 – 95 % (depending on the detector location) of all neutrons hitting the detectors came via the torus ports, meaning that only 5-10 % of the detected neutrons penetrate the tokamak wall. The share of the latter is higher for detectors close to the neutron source, where the neutron spectrum is harder.
Figure 4-2. Relative flagged FC response via ports versus detector position. Neutron source at 0 ° (left) and 22.5 ° (right).
The largest contribution to FC detector response at a certain position is via the nearest port to the detector. The second largest contribution is from the port closest to the neutron source, in our case the east port. For the detectors close to the source (within ± 60 °) these two contributions present approximately 90 % of the total FC response. In more distant detectors, however, contributions from other ports become more important, as their shares are
on the order of 5 % or more. It is interesting to note that ~ 25 % of the neutrons contributing to the west FCs come via the east port. A great majority of these neutrons are scattered from the torus hall wall. If the source is located between the ports, a larger proportion of neutrons scatter around inside the torus, relatively increasing the importance of the port close to the detector.
4.3 Massive items outside the torus
The space around the torus in the torus hall wall is congested by diagnostic equipment, heating facilities, iso-containers, etc. During the calibration there will be two remote handling tents located at the opposite ends of the tokamak. However they will not be present during the JET operation and will be replaced by various plasma diagnostics equipment. Hence it is important to understand and evaluate the effect of large items around the torus on the FC response.
Figure 4-3: Difference in FC response (structure/no structure - 1) due to presence of large structure in front of various ports, as annotated. Neutron source at 0 ° (left) and 22.5 ° (right).
This was examined by an extreme example, i.e. by placing a block of Iron (2 m × 2 m × 2 m) in front of various ports and observing the difference in FC response. The results are presented in Figure 4-3. A block of Fe in front of the east port completely blocks the neutrons hitting the FCs via this port. This conclusion is in very good agreement with [1] showing the relative contribution to the FC via the ports. In Figure 4-2 we can observe that ~90 % of neutrons hitting the FCs in the east octant come via the E port. When blocking the E port with the Fe block the FC response is decreased by almost 90 %. A similar conclusion holds for other detectors as well. Blocking the port closest to the source affects practically all FCs. This kind of information will be used to derive final corrections appropriate to the difference between the conditions at the time of the calibrations and in the plasma conditions after JET restart.
5 CONCLUSIONS
Accurate absolute calibration of external neutron monitors is a very difficult task as it depends on many variables. As the conditions during calibration can be very different from the ones during operation (due to various structures inside and outside the torus) it is essential to thoroughly understand the transport of neutrons from the vacuum vessel to the external detectors. In this paper we examined the major factors affecting the FC response. The main
findings are the following. The highest contribution to a certain FC is via the closest port and the second highest contribution is via the port closest to the neutron source. If the port is blocked by a massive object, the FC response is decreased by approximately the contribution of that port. It was observed that the torus hall wall significantly affects the external monitor response due to back scattering of neutrons. Only 5- 10 % of the neutrons hitting the FC have penetrated the tokamak wall, the others have come via the ports.
In order to thoroughly understand and to correctly interpret the calibration results, more calculations will be performed in the future, focusing on the effect of neutron source shape and spectra, port closure plates, large items inside the vessel (boom and the JET mascot robot), etc. Final corrections will be derived from calculations similar to these, but run within a complete detailed model of the torus and its immediate environment.
It should be noted that this whole process of understanding and improving the knowledge of the neutron yield calibration for JET is of great interest for ITER, where the methods and procedures for calibrating the (different) neutron yield monitors are being developed, but the requirement is for 10 % accuracy in the fusion yield determination, as it is in JET.
ACKNOWLEDGMENTS
This work was supported by EURATOM and carried out within the framework of the European Fusion Development Agreement. The views and opinions expressed herein do not necessarily reflect those of the European Commission.
REFERENCES
1. Laundy Brian J., J.O.N., Numerical study of the calibration factors for the neutron counters in use at the Joint European Torus. Fusion Technology, 1993. 24: p. 150- 160.
2. Johnson, M.G. and et al., Modelling and TOFOR measurements of scattered neutrons at JET. Plasma Physics and Controlled Fusion, 2010. 52(8): p. 085002.
3. Team, X.-M.C., MCNP - A general Monte Carlo N-particle Transport code, Version 5. 2003(LA-UR-03-1987).
4. McLane, V., ENDF-201, ENDF/B-VI Summary documentation supplement I. 1996, National nuclear data center Brookhaven national laboratory: Brookhaven. p. 304.
5. D. Lopez Aldama, A.T., FENDL-2.1 Update of an evaluated nuclear data library for fusion applications. 2004, INTERNATIONAL ATOMIC ENERGY AGENCY:
Vienna. p. 34.
6. Koning A., F.R., Kellett M., Mills R., Henriksson H., Rugama Y., The JEFF-3.1 Nuclear Data Library. 2006, NUCLEAR ENERGY AGENCY ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT: Paris. p. 130.
7. Chadwick, M.B., et al., ENDF/B-VII.0: Next Generation Evaluated Nuclear Data Library for Nuclear Science and Technology. Nuclear Data Sheets, 2006. 107(12): p.
2931-3060.
8. IAEA, International Reactor Dosimetry file 2002 (IRDF-2002), in Technical report series. 2006, International Atomic Energy Agency: Vienna. p. 162.
