Résumé du preprint Irfu-08-431

Irfu-08-431
Neutronic and nuclear post-test analysis of MEGAPIE
S. Panebianco, J.C. David, D. Doré, S. Lemaire, S. Leray, A. Letourneau, F. Mihel-Sendis, A. Prevost, D. Ridikas, G. Stankunas, M. Eid, C. Latge et al
Scope of the work The spallation target is one of the most delicate and technically challenging components of an Accelerator-Driven System (ADS)*. The MEGAPIE experiment has been very important since it consisted in the first experiment to design, manufacture, test, commission and operate a liquid metal spallation target at the MW beam power level. The design phase of the project was supported by a R&D work divided in several areas, and performed by specialists from the corresponding research fields, studying problems related to thermal-hydraulics, structural mechanics, materials issues, neutronic and nuclear assessment. Concerning the latter topics, in the design phase a comprehensive study was performed on neutronic benchmark, target performance, power deposition, radiation damage to structural materials, target activation, gas production, and shielding. A lot of the work performed needed a validation in the irradiation phase, where neutronic and nuclear measurements, and corresponding calculations, could be performed. The goals of the work presented in this report were many: i) characterize the SINQ facility with the MEGAPIE target from the neutronic point of view, by measuring neutron fluxes at various points of the facility; ii) measure the flux inside the spallation neutron source; iii) compare the neutronic performance of MEGAPIE with the ones of the solid targets used routinely at SINQ; iv) measure the delayed neutrons; v) measure the gas release from the target following irradiation; vi) complement each measurement with Monte Carlo calculations, for the purpose of validating the codes used during the design and measurement phases of the project, and if necessary improve the previous calculations; vii) perform target activation calculations of interest to the target disposal and the post-irradiation experiment. Such an ambitious group of tasks could be performed thanks to the deep involvement of experts from many of the institutions part of the MEGAPIE collaboration. Neutronic performance The spallation neutron source of an ADS must provide the additional neutrons needed to drive a subcritical core. The spallation source must be highly efficient given the constraints of the beam current and of the subcritical core. While it is outside the scope of the MEGAPIE experiment to measure the neutron yield of a spallation target, by measuring the neutron fluxes at several points of the target and of the surrounding facility it is possible to characterize the neutronic behavior of the spallation source. Basic measurements consist in monitoring the neutron fluxes at the normal measurement points of the facility. This offers several advantages: standard and very reliable measurement techniques can be used; Monte Carlo codes used for the neutronic simulations can be validated; the neutronic performance of the liquid metal target can be directly compared with more traditional types of spallation targets, such as the solid rod types used at SINQ. However, these measurements have one disadvantage: the measured neutron spectra are mostly thermalized, due to the presence of the large heavy water moderator surrounding the SINQ facility. The traditional flux measurements were therefore complemented with flux measurements inside the spallation target. These challenging measurements were accomplished using micrometric fission chambers inserted in the central rod of the MEGAPIE target. The results presented in this report are of great interest: * A list of acronyms is given in Annex E. 8 - The experimentally observed relative increase of the thermal neutron flux of a factor 1.76 using MEGAPIE, with respect to the solid target used before, is reproduced by calculations. These results are complemented by measurements of epithermal and fast neutrons, discussed in the report. - We are able to reproduce by calculations, with discrepancies within 2 standard deviations, most of the thermal fluxes outside the target. - For the measurements outside the target, the highest discrepancy is found at the NAA, which is the closest measurement point, with calculations overestimating the measurements by 30%. - The discrepancy between measurements and calculations is higher in the central rod, where the calculated values are higher than the measured ones by factors from 2 to 3. The fission chamber measurements are in very good agreement with results from monitor foils inserted in close proximity. The experimental work was accompanied by a high-quality simulation program where the spallation targets and the surrounding SINQ facility were described in great detail, using two state-of-the-art particle transport codes such as FLUKA and MCNPX. The fact that the experimental absolute fluxes at the exit of the target block are correctly simulated indicates that overall, the absolute neutronic performances of the targets are correctly calculated. However, discrepancies are found in close proximity to the beam interaction zone, i.e., in a region where the neutron flux is high, it has a high gradient, and a mixed thermal-epithermal-fast spectrum. It is clearly a great challenge to measure the flux close to the interaction point, and it will be important for future investigations on high power liquid metal targets, to understand the origin of these discrepancies. A factor 2-3 between measurements and calculations can be very important in that part of the target, since it is the most subject to radiation damage and to power deposition, with related thermal-hydraulics and structural mechanics issues. Delayed neutrons Another aspect of the neutron flux study that needs to be considered is the delayed neutron flux. DNs are obviously important in reactors, but they can also constitute a safety issue in highpower liquid metal targets. The reason is that, while in the spallation zone the DN flux is negligible compared to the prompt neutron flux, in other areas of the target loop, where the prompt neutrons are shielded, their contribution might be dominant, both in terms of absolute flux and of energy spectrum. That can be important for instance for ancillary components, which must be qualified also to withstand potentially high DN fluxes. The results from the measurements performed during the MEGAPIE start up indeed confirm that at the top of the MEGAPIE target, the absolute DN flux is comparable to the prompt neutron flux. Gas production The production and release of volatile elements during irradiation is a key safety issue in an accelerator-driven system. One of the main disadvantages of an ADS with respect to a fast reactor system is in the amount of volatile elements ending in the cover gas system (CGS), due to the fact that the coolant is directly irradiated by a proton beam, and a large amount of gas is generated by spallation reactions. This requires special care and makes the handling of the gas more complicated. A large program of calculations and experiments was performed in the frame of the MEGAPIE project to assess the volatility of key elements such as noble gases, mercury, and polonium in the MEGAPIE configuration. In the irradiation phase, absolute amounts of released noble gases and of Hg and Au isotopes were determined by the g spectroscopy measurements from the fresh gas sampling made after 2 days of operation. We found that only a 9 fraction of the noble gases produced in the LBE was released into the expansion volume after 2 days of operation. The amount is about 1% of the total, calculated amount. This fraction is nearly the same for Ar, Kr and Xe isotopes, indicating a similar release mechanism for all the noble gas elements. We know from previous measurements at ISOLDE that noble gases diffuse slowly in LBE targets and the diffusion time decreases as the temperature increases. As expected, also the release of Hg was a small fraction of the total, while only traces of Po isotopes were detected from the gas samples, coming presumably from decay of the parent At isotopes. A comparison was performed also with expected release rates during the regular gas samplings. In this case the amount of calculated radioisotopes are within a factor of 3 to the experimental values, indicating, as expected, that the noble gas release is more complete after one month of operation or more. Unfortunately it was not possible to measure directly hydrogen and helium isotopes. Only indirect measurements, from the pressure in the cover gas, were performed and the calculated pressures compared fairly well with the measured ones. Target activation A large amount of nuclear calculations has been performed since the beginning of the MEGAPIE project, and even though they are not part of the post irradiation analysis, they are presented here. The main need for activation calculations in the post-test analysis phase was to determine the activation of the LBE and of the structural materials, for the target decommissioning after irradiation. The main tools were the codes FLUKA, MCNPX, and associated evolution codes, and the SNT code. A large validation and development work has been performed in recent years for these codes. The results from the different calculations were compared for the LBE and structural material activation, showing an overall good agreement for the LBE, with the exception of specific isotopes: isomers, more difficult to calculated, and tritium, emitted or not according to the spallation model used and more or less in good agreement with the few available experimental data. The results for the activation of the target lower structure show some more important differences between FLUKA and MCNPX, presumably coming from the thermal and epithermal region.