Neutron Transport and Material Activation in a Power Plant Based on the HCLL Blanket Concept

Transcription

1 Neutron Transport and Material Activation in a Power Plant Based on the HCLL Blanket Concept R Pampin 1,2, PJ Karditsas 2 and NP Taylor 2 1 The University of Birmingham, School of Physics and Astronomy, Edgbaston, Birmingham B15 2TT, UK 2 EURATOM/UKAEA Fusion Association, Culham Science Centre, Abingdon, Oxon OX14 3DB, UK Corresponding author: R Pampin, raul.pampin@ukaea.org.uk Abstract. The conclusions of the European Power Plant Conceptual Study (PPCS) identified a further near-term concept, to add to the four plants analysed, whose promising potential performance and development as a test module for ITER triggered its evaluation for power production. This is the helium-cooled lithium-lead blanket concept (HCLL). Following the PPCS methodology, a power plant based on this blanket has been conceptually designed, analysed and optimised. Neutron transport and material activation analyses performed for the safety and environmental assessment of the plant, using the purpose-built HERCULES system, are described together with a comprehensive estimation and categorisation of the active waste generated. This shows features similar to other PPCS models, in particular that no permanent disposal waste (PDW) results from the operation and decommissioning of such a power plant, provided low-activation steel Eurofer is used. 1. Introduction The PPCS [1] analysed the performance of four power plants based on magnetically confined burning plasmas: these spanned tokamak physics assumptions and technology concepts ranging from near-term to very advanced, all delivering high grade heat for power production. A fifth power plant, based on the HCLL blanket and named plant model AB (PM-AB), is currently being studied. Its physics assumptions are similar to those already explored in the near-term PPCS Models A and B. The HCLL concept derives from a European ITER test blanket module proposal, using LiPb breeder and helium coolant and achieving sufficient tritium production, energy multiplication and high thermodynamic efficiency. This paper presents the calculations performed for the safety and environmental assessment of this model, and in particular for obtaining levels of activation, decay heat and other radiological quantities enabling the estimation of: (a) radioactive material inventory, (b) temperature excursions during bounding, decay-heat driven accidents, and (c) operational safety parameters. They were performed with 3D computational models of the plant built with the HERCULES code [2]. The models were used in MCNP, [3], to determine the neutron spectra throughout the structures surrounding the plasma, and in the FISPACT code coupled to the EAF nuclear data libraries, [4,5], to perform the calculation of the neutron activation inventory and related quantities. 2. Description of Plant Model AB The HCLL blanket concept is entirely cooled by helium at 8 MPa (T in ~300 o C, T out ~500 o C), and uses eutectic LiPb to breed tritium in-situ by circulating it slowly through the blanket and detritiation loops in order to generate and extract this isotope while avoiding MHD interaction with the magnetic fields. It is divided into 180 modules mounted on the vacuum vessel and completely surrounding the plasma, [6]. It assumes a 2mm tungsten armour coating the first wall (FW), due to concerns arising from plasma-surface interactions. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

2 The overall layout consists of armoured first wall, HCLL blanket, high and low temperature shields (HTS and LTS), divertor, vacuum vessel (VV) and magnets. The HTS is a replaceable component attached to the blanket modules, and made of helium-cooled steel. The LTS is a lifetime component attached to the VV and made of steel structure, tungsten carbide for neutron shielding, and water coolant. The VV is made of steel and small amounts of boron, and is also water-cooled. The magnets considered in this study include 18 superconducting toroidal field coils (TFC) and a central solenoid (CS), both derived from the ITER designs. RAFM Eurofer grade steel is the structural material throughout the whole plant, except for the VV and TFC where conventional, grade 316SS is used. Both the HEMP and HETS highperformance, gas-cooled concepts are considered for PM-AB divertor, due to compatibility with the blanket cooling system. Neutronic and activation analyses of this structure will be subject of detailed assessment elsewhere. For the purpose of this study, only a very simplified representation was used in order to provide an approximation to the overall behaviour: both of them can be generally considered as a helium-cooled steel structure and heat sink, upon which tungsten target plates are mounted. 3. Computational Modelling An MCNP model was developed in order to perform neutron transport analyses, which was to be used later for the activation and accident consequence calculations as well. Table 1 summarises the machine and plasma parameters, found after economic optimisation, that were supplied to HERCULES for the generation of the model. This computer programme assisted in the automation of geometry and MCNP and FISPACT input file generation. The radial build of PM-AB was approximated within the constraints of the code: see Table 2 providing details. Figure 1 shows a poloidal cross section of the model generated for the MCNP calculations, including the poloidal angle segmentation and plasma shape, and a detail of the outboard midplane radial build in which the different components can be identified. The model consists of 11 layers, differentiating between inboard and outboard, and 16 poloidal sectors. Gap layers are automatically added between the four outer ones. Each component was FW breeder HTS LTS VV TFC manifold Fig. 1: PM-AB MCNP model generated by HERCULES showing the poloidal angle segmentation and plasma shape, left, and detail of the outboard midplane, right. Dark cells represent the divertor. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

3 TABLE 1: PLANT MODEL AB PARAMETERS Parameter PM-AB Fusion Power (GW) Major Radius (m) 9.56 Aspect Ratio 3.0 Scrape-off layer thickness (m) 0.15 Machine elongation 1.9 Machine triangularity 0.4 Plasma ion core temperature (kev) 53.8 Plasma elongation 1.7 Plasma triangularity 0.27 Plasma peaking factor 1.7 fitted to a single radial layer except the blanket breeder region which, as seen in Table 2, consisted of four of these in order to increase the detail of the modelling. The divertor of PM- AB was approximated into a semi-void steel structure with the plasma-facing tiles represented by the first two layers at 100% volume of tungsten. The cryostat is cylindrical and made of conventional steel, and the CS has an inner radius of 2.4m and the same material composition as the TF coils. 3.1 Neutron Source The plasma neutron source in the MCNP model was a typical isotropic, 14.1 MeV gaussian D-T fusion source, represented using the parameters listed in Table 1. The shape is ellipticaltriangular, assuming magnetic flux surfaces as surfaces of constant neutron emission. The radial intensity is controlled with a peaking factor. The optimised fusion power for PM-AB is 4.02 GW, corresponding to a total neutron yield of n/s. TABLE 2: FINAL PPCS PM-AB RADIAL BUILD UP Outboard inboard layer component thickness (m) cumul. (m) thickness (m) cumul. (m) material composition * 1 TF coil ITER coil mixture 2 Vacuum vessel SS + H 2 O + boron (61.4/37.0/1.6 %) 3 LT shield WC + Eurofer + H 2 O (65/10/25 %) 4 HT shield Eurofer + He (50/50%) 5 Blanket Eurofer + He (50/50%) manifold 6-9 Blanket breeder LiPb + Eurofer + He (80/10/10 %) 10 Blanket FW Eurofer + He (70/30%) 11 FW armour W * Volume fractions in %. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

4 1 2 3 full power (a) (b) (c) (d) zero power plant shutdown Fig. 2: Proposed PPCS operational scheme: (a) 2.5 years at full power operation, (b) two months for maintenance and divertor replacement, (c) another 2.5 years at full power, and (d) ten months for maintenance, divertor and blanket replacement. The cycle is repeated five times, obviously excluding the last replacement outage. This scheme results in a planned availability of 85.7%. 3.2 Material Specifications The material elemental compositions applied in the neutron transport calculations and those in the activation analyses varied slightly. Only the more abundant elements were considered in the former, whereas for the latter small traces of impurities, not relevant for neutron transport but important in the activation behaviour of the materials, were included. These specifications were the result of a series of studies dedicated to the optimisation of those materials regarding activation properties and feasibility concerns, including a conservative list of impurities [7]. Only the TFC and CS material composition was obtained from a different source, being a mixture of epoxy, copper, conventional steel, Nb 3 Sn superconductor and helium coolant as in the ITER coils, at 5.66 g/cc density, [8]. Natural isotopic abundance was assumed throughout the model except for the lithium in the breeder, assumed to be enriched to 90% 6 Li. 3.3 Irradiation Histories Irradiation histories to be implemented in FISPACT were determined from available information on PPCS model AB design and PPCS maintenance scheme, [9], sketched in Figure 2, assuming all components are at the end of their operational life. According to this, the irradiation histories of the different structures could be classified in just three different schemes, shown in the Figure: (1) for components replaced after a 2.5 years irradiation cycle (divertor), (2) for components replaced after a 5 years irradiation cycle (blanket, including FW and its armour, and HTS), and (3) for lifetime components (LTS, VV, TFC and CS). 4. Results and Discussion 4.1 Neutron Transport Analysis The model generated by HERCULES was used in MCNP 4C3: for efficiency of the computation an importance cell map was implemented by splitting and Russian roulette; no other variance reduction techniques were employed. Nuclear data tables were taken from the ENDF/B-VI evaluated library. Neutron energy spectra were calculated in all non-void cells of the model, and binned following the 175-group Vitamin-J format. An illustrative sample of the results is presented here. Figure 3 shows energy spectrum histograms of the neutron flux in cells located at different radial locations of the outboard midplane of the model, and Figure 4 the poloidal variation of the neutron flux at those same radial locations. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

5 1.E+19 1.E+18 1.E+17 flux per unit lethargy 1.E+16 1.E+15 1.E+14 FW armour 1.E+13 FW breeder front 1.E+12 breeder back 1.E+11 HT shield LT shield 1.E+10 VV TFC 1.E+09 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 energy (MeV) Fig. 3: Neutron spectra (flux per unit lethargy, n/m 2 s) at different radial locations of the outboard midplane of PM-AB. Spectra show normal features. In the breeder region moderation is poor and (n,2n) reactions in lead dominate the neutron behaviour: great similarity to that obtained for the previous PPCS models using DCLL and SCLL blankets is observed, [2]. The depression in the breeder flux (blue curves) at ~250 kev is due to a resonance in the 6 Li(n,T) 4 He reaction cross section. Iron giant resonances are observed nearly everywhere due to the presence of steel throughout the plant; tungsten ones also where this material is present (e.g. FW armour and LT shield at ~20 ev). Much more moderation and attenuation is achieved by means of the tungsten carbide in the low temperature shield and boron in the vacuum vessel. Poloidal variation is also very similar to that observed in previous PPCS models: in the first, plasma facing layer the flux is maximum at the outboard midplane segments and drops off towards the top and bottom of the machine, before rising again in the inboard midplane. This dependence is expected due to the neutron source being centred in the midplane, and the flux in these layers being mostly unscattered. total neutron flux (n/m2s) 1.E+20 1.E+19 1.E+18 1.E+17 1.E+16 1.E+15 1.E+14 1.E+13 1.E+12 1.E+11 1.E+10 1.E+09 1.E+08 1.E+07 1.E poloidal angle total neutron flux (n/m2s) 1.50E E E E E E+18 FW armour FW breeder front breeder back HT shield LT shield VV TF coils 0.00E poloidal angle Fig. 4: Poloidal variation of the total neutron flux at different radial locations of the outboard midplane of PM-AB, left, and details of this variation in the first layers, right. The two leftmost and rightmost points in all layers correspond to divertor segments. Note: the legend shown in the right hand side applies to both graphs. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

6 TABLE 3: HERCULES NEUTRON TRANSPOR RESULTS AND COMPARISON WITH DETAILED PPCS ANALYSES Parameter HERCULES Detailed model Average neutron wall load (MW/m 2 ) FW avg. neutron wall load (MW/m 2 ) Divertor avg. neutron wall load (MW/m 2 ) Total nuclear heating (MW) 3924 * 3987 Energy multiplication Average nuclear power density (MW/m 3 ) Total TBR * Of which FW+armour: 11.1%, breeder+manifold: 71.2%, divertor: 13.8%, HTS: 1.3%, LTS: 2.6% The maxima pattern is gradually reversed in inner layers, shifting towards the inboard due to the lower attenuation here (less thickness). The further away from the plasma, the more different the thickness and thus the more exaggerated the poloidal variation. HERCULES neutron transport module has capabilities extending beyond the calculation of spectra for neutron-induced activation analyses during safety assessments, [10]. The input file it produced included material and tally definitions that, after MCNP neutron-photon analysis, allowed the post-processing module to calculate parameters of crucial importance in fusion technology design. Radial and poloidal variations, as well as total and/or average values of tritium generation, energy deposition, neutron wall load and radiation damage rates to materials were so obtained. Results presented here allow comparison with the more extensive neutronic analysis of PM-AB, [6]. Table 3 shows a summary of these results and comparison. All values are within 2% of each other. Figure 5 shows midplane radial profiles of damage rate to the structural material, and Figure 6 tritium breeding ratio (TBR) poloidal variation in different radial layers of breeder. Total and inboard tritium generation seem appropriate, whilst inefficient at the outboard rear of the blanket. It is acknowledged that the detailed model had slightly thicker inboard and outboard breeder regions, and that the HERCULES one was more homogeneous and did not account for poloidal/toroidal gaps. radiation damage rate (dpa/year) OUTBOARD: blanket HTS LTS outboard inboard INBOARD: blanket HTS LTS VV radial thickness (cm) TBR / kg of breeder layer 9 layer 8 layer 7 layer poloidal angle Fig.5: Radial profiles at inboard and outboard midplane of the radiation damage rate (dpa/year) to the structural material. Fig.6: Poloidal variation of the tritium breeding ratio for the different layers of breeder in the radial build up. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

7 The only exception to the otherwise good statistics of these Monte Carlo calculations, achieved through the combined use of importance mapping and parallel computing, were cells in the outboard TFC of the model. Transport of neutrons to this region was found difficult to achieve, and several cells had total tallies with 100% relative error. A conservative, constant neutron flux value and spectrum throughout the outboard TFC was assumed and used in the activation calculations (see constant flux from segments 2 to 8 in Figure 4). This was expected to produce slight overestimation of the otherwise insignificant activation of this structure. 4.2 Activation Analysis MCNP produced an output file containing the neutron spectra of every cell in PM-AB, suitable for activation calculations to be performed using HERCULES coupled to the inventory code FISPACT and the EAF data libraries (ver 2003). Presented here is an illustrative sample of the vast amount of radiological information obtained, which included: (i) inventories of 1917 nuclide species, (ii) total levels of specific activity, decay heat, contact dose rate, ingestion and inhalation doses, (iii) clearance index and (iv) uncertainties of all the above quantities. All these were produced for every cell in the model and at a series of predetermined time-steps, from plant shutdown up to years later. This provides the basis for subsequent bounding accident consequence analyses, but also for the estimation and categorisation of the radioactive waste generated during the operation and decommissioning of such a power plant reported in section 4.3. Time histories of the specific activity, specific decay heat and contact dose rate for different radial locations can be found in Figures 7, 9 and 10. Poloidal variation of the former at time zero is shown in Figure 8. It should be noticed that the two rightmost and leftmost points in the poloidal variation correspond to the divertor region. Some of the patterns observed in the neutron transport results are repeated here. In general, activation is higher where the neutron flux is higher: at the outboard midplane of the first layers, dropping off towards top and bottom and rising again in the inboard midplane, reversing the trend in inner layers and shifting it towards the inboard. specific activity (Bq/kg) 1.E+16 1.E+15 1.E+14 1.E+13 1.E+12 1.E+11 1.E+10 L hour L day L month L year 1.E+09 1.E FW armour 3- breeder front 1.E breeder back 5- LT shield 1.E VV 1.E TF coils 2- FW 1.E+04 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 time (years) Fig.7: Specific activity histories at different radial locations in the outboard midplane. specific activity (Bq/kg) 1.E+16 1.E+15 1.E+14 1.E+13 1.E+12 1.E+11 1.E+10 1.E+09 1.E+08 1.E+07 FW armour FW 1.E+06 breeder front breeder back 1.E+05 HT shield LT shield VV TF coils 1.E poloidal angle Fig.8: Specific activity poloidal variation at different radial locations, t=0. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

8 Decay patterns show no anomalous features. Activity in steel, which can be seen uncovered by other materials in the FW, is dominated by 55 Fe (t 1/2 =2.7y). The main pathways are 56 Fe(n,2n) 55 Fe for hard, poorly moderated spectra (like those in the FW), and 54 Fe(n,γ) 55 Fe for soft, moderated spectra (like those in the rear blanket and HTS). Of more relevance for subsequent safety analyses than the activity, however, is the heat generated through radioactive decay, particularly in the structural material. The main nuclides responsible for this decay heat in Eurofer located at the FW and rear of the blanket (manifold) are shown in Figures 11 and 12. It can be noticed that for hard spectra (FW) decay heat during the usual accidental time scale (1 to 100 days, dotted lines) is dominated by 54 Mn (β +, 312.3d) generated mainly through 54 Fe(n,p) 54 Mn reactions, whereas for soft spectra (manifold) the dominant nuclide during the same time scale is 182 Ta (β -, 114.7d) generated via 181 Ta(n,γ) 182 Ta. Contact dose rate is found to have similar dominant nuclides. Activity in the breeder is vastly dominated by tritium (12.33y), whose main pathway is 6 Li(n,T) 4 He, although also 7 Li(n,n T) 4 He makes a small contribution. specific decay heat (kw/kg) 1.E+00 1.E-01 1.E-02 1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 1.E-08 1.E-09 L hour L day L month L year 1- FW armour 3- breeder front 4- breeder back 5- LT shield 6- VV 7- TF coils 2- FW 1.E-10 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 time (years) Fig.9: Specific decay heat histories at different radial locations in the outboard midplane. contact dose rate (Sv/h) 1.E+06 1.E+05 1.E+04 1.E+03 1.E+02 1.E+01 1.E+00 1.E-01 1.E-02 1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 L hour L day L month L year 1- FW armour 3- breeder front 4- breeder back 5- LT shield 6- VV 7- TF coils 2- FW 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 time (years) Fig.10: Contact dose rate histories at different radial locations in the outboard midplane. 100 Mn-56 Mn-54 Fe-55 Co-60 C Ta-182 Co-60 Nb-94 Nb-94 W-187 C-14 percentage decay power 10 Ta-182 Cr-51 Ar-39 percentage decay power 10 Mn-56 As-76 Fe-59 Mn-54 W-185 Fe-55 W-185 Mn-53 Sb time (years) Fig.11: Dominant nuclides responsible for the decay heat generation of Eurofer in the FW time (years) Fig.12: Dominant nuclides for the decay heat generation of Eurofer in the manifold. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

9 Mass (tonnes) Decay time (y) total NAW total SRM total CRM total PDW 0 Fig. 13: PM-AB masses (tons) of the material from the various regions of the plant after 50 and 100 years of cooling. All replacements of the same component are included. 4.3 Assessment of Activated Material Based on the FISPACT results for clearance index, contact dose and decay heat rates, it was possible to categorise the active material in the plant according to the criteria adopted in the PPCS programme, [9]. These define categories for non-active material, simple and complex recycling and permanent disposal waste. A detailed breakdown into such categories of PM- AB material, and comparison with the other PPCS models, can be found in [11]. Figure 13 summarises the total masses of each category at 50 and 100 years after shutdown, and Figure 14 shows a poloidal section in which a colour scheme identifies the category the different components lie in after 100 years of cooling. The total mass of active material generated in PM-AB over its lifetime is tons, including all replacements of the same component. On the basis for recycling potential used in the PPCS, 50 years after plant shutdown there are tonnes of material requiring permanent disposal. The situation after 100 years is greatly improved, and no PDW exists anymore: most of the material is NAW and SRM, and CRM SRM SRM - hands on NAW Fig. 14: Poloidal cross section of the plant showing the category the different components fall in after 100 years of cooling. Note that material showing potential for hands-on recycling operation (<10 µsv/h contact dose rate) is also shown. R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July

10 only ~13% is CRM. All outboard TFC and nearly all outboard VV sectors are NAW, and the remaining parts of the latter show potential for hands-on recycling. Regarding other PPCS targets, outboard TFC sectors seem to achieve clearance after only 50 years, but hands-on operation for in-vessel components is not achieved at all, even after 100 years. As previously noticed for other PPCS models, tungsten armouring affects only mildly the results and in no case generates PDW. It does affect, however, safety issues arising during postulated accidents and maintenance operations due to high tungsten early gamma dose, [12]. 5. Conclusions The neutron transport and activation analyses performed using the HERCULES system for the safety and environmental evaluation of a power plant based on the HCLL blanket are reported. These have produced neutron spectra throughout the tokamak structures, and a comprehensive estimation and categorisation of the active inventory, showing features similar to previous PPCS models. In particular, it is found that no material activated during the operation of such a power plant would be categorised as permanent disposal waste 100 years after the end of life, provided low-activation steel Eurofer is used. Total values and radial/poloidal variations of several nuclear parameters have also been obtained. Comparison with results of more detailed neutronic assessments show the overall consistency of the code. ACKNOWLEDGMENTS This work was jointly funded by the UK Engineering and Physical Sciences Research Council and EURATOM REFERENCES [1] MARBACH, G., COOK, I. and MAISONNIER, D., The EU Power Plant Conceptual Study, Fusion Engineering and Design (2002), 1-9. [2] PAMPIN, R., Fusion power: safety and environmental analysis using integrated, threedimensional computer modelling, PhD Thesis, University of Birmingham, UK (2004). [3] BRIESMEISTER, J.F., MCNP A general Monte Carlo N-Particle transport code, version 4B, LA M [4] FORREST, R.A., FISPACT-2003 user manual, UKAEA FUS 485, January [5] FORREST, R.A., et al., The European Activation File: EAF-2003 cross section library, UKAEA FUS 486, December [6] LIPUMA, A., et al., Breeder blanket design and systems integration for a heliumcooled lithium-lead fusion power plant, to appear in Fusion Engineering and Design (Proc.of the 7th ISFNT, Tokyo, May 2005). [7] FORTY, C.B.A., Compositional optimisation of SEAFP-2 materials, UKAEA report, SEAFP2/4.1/UKAEA/4, August [8] FISCHER, U., private communication, 7 October [9] MAISIONNIER, D., et al. for the PPCS team, A conceptual study of commercial fusion power plants: final report of the European Power Plant Conceptual Study (PPCS), EFDA report, EFDA-RP-RE-5.0, September [10] PAMPIN, R., et al., Fusion power plant nuclear performance analysis using the HERCULES code, to appear in Fus. Eng. and Des. (Proc. of the 7th ISFNT, 2005). [11] FORREST, R.A., et al., Categorisation of activation waste from PPCS model power plants, this meeting. [12] TAYLOR, N.P., et al., Activation properties of tungsten as a first wall protection in fusion power plants, to appear in Fus. Eng. and Des. (Proc. of the 7th ISFNT, 2005). R Pampin IAEA Technical Meeting on First Generation of Fusion Power Plants, Vienna, July