A Neutron Poison Tritium Breeding Controller Applied to a WCCB Fusion Reactor Model

Transcription

1 CCFE-PR(14)22 L.W.G. Morgan and L.W. Packer A Neutron Poison Tritium Breeding Controller Applied to a WCCB Fusion Reactor Model

2 Enquiries about copyright and reproduction should in the first instance be addressed to the Culham Publications Officer, Culham Centre for Fusion Energy (CCFE), Library, Culham Science Centre, Abingdon, Oxfordshire, OX14 3DB, UK. The United Kingdom Atomic Energy Authority is the copyright holder.

3 A Neutron Poison Tritium Breeding Controller Applied to a WCCB Fusion Reactor Model L.W.G. Morgan 1, L.W. Packer 1 1 EURATOM/CCFE Association, Culham Science Centre, OX14 3DB Abingdon, UK. The following article appeared in Proceedings of the 11th International Symposium on Fusion Nuclear Technology (ISFNT-11) Barcelona, Spain, September, Fusion Engineering and Design, Vol.89, Issues 7-8, October 2014, pp Further reproduction distribution of this paper is subject to the journal publication rules.

4 .

5 2014 UNITED KINGDOM ATOMIC ENERGY AUTHORITY The following article appeared in Proceedings of the 11th International Symposium on Fusion Nuclear Technology (ISFNT-11) Barcelona, Spain, September, Fusion Engineering and Design, Vol.89, Issues 7-8, October 2014, pp A neutron poison tritium breeding controller applied to a water cooled fusion reactor model Morgan L W G, Packer L W The Version of Record is available online at /j.fusengdes

6 A neutron poison tritium breeding controller applied to a WCCB fusion reactor model L. W. G. Morgan a, L. W. Packer a a EURATOM/CCFE Fusion Association Culham Science Centre, Abingdon, Oxon, OX14 3DB, UK. Abstract The generation of tritium in sufficient quantities is an absolute requirement for a next step fusion device such as DEMO due to the scarcity of tritium sources. A number of methods have been proposed in order to meet this requirement, however a lithium-based tritium breeding blanket that surrounds the fusion plasma is widely considered to be the most suitable solution. Although the production of sufficient quantities of tritium will be one of the main challenges for DEMO, within an energy economy featuring several fusion power plants the active control of tritium production may be required in order to manage surplus tritium inventories at power plant sites. The primary reason for controlling the tritium inventory in such an economy would therefore be to minimise the risk and storage costs associated with large quantities of surplus tritium. The composition of liquid breeding materials could be potentially be adjusted on-line in order to control the amount of impurities and the amount of tritium being produced, however the composition of solid breeders could not be easily changed with a short replacement time. In order to ensure that enough tritium will be produced in a reactor which contains a solid tritium breeder, over the reactor s lifetime, the tritium breeding rate at the beginning of its lifetime is relatively high and reduces over time. This causes a large surplus tritium inventory to build up until approximately halfway through the lifetime of the blanket, when the inventory begins to decrease. This surplus tritium inventory could exceed several tens of kilograms of tritium, impacting on possible safety and licensing conditions that may exist. This paper describes a possible solution to the surplus tritium inventory problem that includes neutron poison injection, which is managed with a tritium breeding controller. A PID controller is used to inject neutron absorbing compounds into the water coolant, depending on the difference between the required tritium excess inventory and the measured tritium excess inventory. The compounds effectively reduce the amount of low energy neutrons available to react with lithium compounds, thus reducing the tritium breeding ratio. This controller reduces the amount of tritium being produced at the start of the reactor s lifetime and increases the rate of tritium production towards the end of it lifetime. Thus, a relatively stable tritium production level may be maintained, allowing the control system to minimize the stored tritium with obvious safety benefits. Previous tritium breeding studies have shown that coupled neutron-transport and burn-up is required in order to accurately predict the tritium self-sufficiency. Hence, the FATI code (Fusion Activation and Transport Interface) will be used to perform the tritium breeding and controller calculations. FATI couples MCNP and FISPACT-II in a cyclic manner and implements controller mechanisms between transport and burn-up steps. Results are presented from the FATI code applied to a simplified HCCB DEMO conceptual model. Keywords: monte-carlo, burn-up, depletion, tritium breeding, tritium self-sufficiency, DEMO 1. Introduction Tritium self-sufficiency is an absolute requirement for DEMO and commercial fusion power plants. Thus, the modelling of tritium breeding blankets must ensure the blanket is able to acheieve a large tritium breeding margin with a high degree of certainty. However, the over-production of tritium is also of concern due to safety considerations, tritium storage issues and tritium licensing [1]. The risk associated with the under-production of tritium is greater than the over-production, however this issue still needs to be addressed. El-Guebaly [1] points out that the uncertainty related to the TBR can be broken down to nuclear data (6-10%), modelling (3-7%), design (0-3%) and tritium consumption (1-2%). An uncertainty of just 1% can translate to to an over-/under-production of more address: Lee.Morgan@CCFE.ac.uk () Tel.: (+44) () than 1 kg/year of tritium, which could pose a problem for self-sufficiency or licensing/storage of tritium. Thus, the overproduction of tritium is likely to become more probable as the uncertainty on the need to over-produce it is reduced. The production issue can be resolved with the online control of tritium. Studies of the online control of tritium production has concentrated on LiPb liquid blankets [2, 3], however, liquid metal breeding blankets are technologicially more advanced than solid types breeders due to the high temperature conitions. 2. Reactor Model The use of liquid metal as a coolant/breeding material within fusion blankets is highly likely due to the reduced radioactive waste, increased plant availability and ease of tritium breeding control when compared with solid breeders. To date, solid breeders have been considered to have a predetermined, unchangeable tritium production scheme, with a study concluding Preprint submitted to Fusion Engineering and Design August 20, 2013

7 Component Radial Density Material composition depth (cm) (g/cm 3 ) Void First wall CuCrZr (0.05), Eurofer (0.9), Helium (0.05) Breeding blanket Lithium-Orthosilicate(0.63), Berylium (0.07), Water (0.3) Back plate Eurofer (0.8), Helium ( 0.2) Vessel Helium (0.2), Inconel-718 (0.8) TF coils LN (0.7), Helium (0.2), Inconel-718 (0.1) Table 1: WCCB model (figure 2) dimensions and composition that the online control of tritium production within solid breeders not being foreseeable [1]. As a result of the radiotoxic nuclides produced in liquid lithium-lead blankets, a solution is sought that combines the tritium breeding controlability of liquid breeders with the low activation of solid breeders. Thus, this paper describes a method, based on neutron poisons, which may enable the online control of tritium production within solid breeder blankets such as helium cooled pebble bed (HCPB) and water cooled ceramic breeder (WCCB). Solid breeder blankets consist of tritium breeder and neutron multiplier pebbles that are fluid cooled. The breeding pebbles are sub-millimeter sized and are composed of a lithium ceramic, such as Li 4 S O 4 or Li 2 TiO 3, and the neutron multiplier is usually berylium. A cooling fluid, such as helium or water, also acts to purge the tritium from the blanket. This work focuses on the Water Cooled Ceramic Breeder (WCCB), which has been the focus of study of the Japanese fusion program [4, 5]. A schematic diagram and composition of the WCCB model is shown in figure 2 and table Neutron poison For this study, boric acid (H 3 BO 3 ) injection into water coolant has been chosen to reduce TBR in a WCCB reactor. Boron is widely used in the fission industry [6] and is readily available and relatively cheap. 7 Li is produced as a result of the neutron capture by 1 0B. While 7 Li contributes to the production of tritium, its (n,n t) cross-section (and associated reaction rate) is less than the 10 B(n, α γ) 7 Li cross-section, which will reduce the production of tritium. Boron trifluoride ( 10 BF 3 ), a biohazardous gas, could be used to control the TBR in a HCPB blanket and is one a only a few gases that are suitable as a boron neutron poison. The use of 10 BF 3 would not be ideal, however it will be confined in an environment that contains tritium, which is biohazardous and radiotoxic. 3. Computational method Figure 1: Simple spherical WCCB model. Compositions and dimensions are shown in table 1. (Not to scale) Design A simple water cooled ceramic breeder model, consisting of a 14.1 MeV point neutron source and a spherical blanket, is to be used in conjuction with the neutron poison based tritium controller. The key components of this simple model are the 2 GW point fusion source and a single spherical breeding module comprised of lithium orthosilicate (Li 4 S O 4, natural 6 Li enrichment), and berylium pebbles, which are cooled with water. Recent studies [7, 8] have shown that Cyclic Coupling of Radiation-Transport and Burn-up (CCRTB) has a significant effect on the tritium self-sufficiency time (TSST). Earlier work on tritium self-sufficiency and tritium production have comprehensively outlined uncertainties and potential isues relating to tritium breeding. However, these methods have either been analytical [9], based on non-time-dependent burn-up [10, 11, 12] or static radiation-transport and burn-up [13]. In some cases the TSST calculated by CCRTB is more than double that predicted by the non-crtb calculations, which assume a constant neutron flux throughout the lifetime of the blanket. This is due to the decreasing amount of 6 Li in the blanket, resulting in fewer thermal neutrons being absorbed, which leads to the growth of the thermal neutron population as a function of time. Therefore,the use of CCRTB is recomended for time-dependent tritium breeding calculations. The coupling of radiation transport and burn-up has been peformed by many codes and for many years in the fission industry (MCODE, MOCUP, MON- TEBURNS, VESTA), however the use of such codes is less common within the fusion neutronics community. The CRTB code utilised in this study is FATI (Fusion Activation and transport Interface), which currently interfaces MC- NPX [14] with FISPACT [15]. The intended primary application of FATI is the simulation of nuclide burn-up within fusion blankets. 2

8 3.1. Neutron poison controller The PID ( Proportional Integral Derivative) controller is one of the first control strategies to be implemented and is a commonly used feedback method [16, 17]. The output of the PID controller, u(t), is given by: u(t) = K P e(t) + K D de(t) dt + K I t 0 e(t).dt where e(t) is the error, defined by the difference between the desired output, y(t), and the controller output, u(t), e(t) = u(t) y(t). The parameters K P, K I and K D are user defined constants and the desired state to be controlled is the tritium inventory. In this study, the control variable is boron concentration, B. A conventional controller varies the control parameter with an infinite scale; however the control parameter, B, is defined in the interval [ 0, Bmax] ND ( see figure 2). This issue is resolved by introducing a transfer function, F(B ), which maps the output of the PID controller, B new to the interval [ 0, Bmax] ND, creating a variable Bnew, ND which represents the new concentration of boron. The transfer function, shown graphically in figure 2, is analytically defined as: ( ( F = Bmax ND ( 1 exp B ) 2 new /σ )) where σ is the standard deviation of the Gaussian function. Hence, the procedure for updating the boron concentration, C, for each cell (or group of cells) is as follows: 1. Map to current boron number density, B ND old, to the infinite range, B old using the inverse of the transfer function. ]) 1/2 B old = F 1 (B ND old ( σln [1 ) = BND old Bmax ND 2. Calculate to increment in Cold using PID theory: Let e = T i target T i measured B de t = K P E + K D dt + K I E.dt 0 where K P, K I and K D are PID controller control constants. B new = B old + B 3. Map the new ratio to the [0,1] interval ( ( Bnew ND = Bmax ND ( 1 exp B ) 2 new /σ )) This procedure (steps 1-3) is illustrated in figure 2. 3 Figure 2: Mapping of PID output to boron concentration. Figure 3: Surplus tritium inventory, for the WCCB spherical shell model (figure 2), over an operational period of 365 days. Results are shown with and without a boron neutron poison tritium controller. 4. Results Figure 3 shows the tritium surplus inventories, for the WCCB DEMO model shown in figure 2, over an operational period of year, for both a tritium-controlled and a non-tritium-controlled blanket. In order to build the tritium surplus inventory to the required level in a controlled manner, independent of the initial TBR, a linear ramp was used to define the tritium target inventory over time. The tritium surplus target was designed to reach the maximum level of 2 kg of surplus tritium over a period of 120 days. The aim of the tritium controller is to ensure the actual tritium surplus mimics the target. At time = 0, the TBR of both controlled and uncontrolled models, is 1.09 causing the tritium surplus to raise above the taget value (this is evident from the uncontrolled blanket model where the tritium surplus rises very quickly). Hence, the tritium controller increases the amount of boron in the blanket to slow tritium production. Beyond the initial tritium surplus ramping, in the region of target constant tritium surplus the oscillations in the actual tritium surplus attenuates over time as aresult of the differential and integral functions within the tritium controller. The correlation between TBR and boron conentration is evident in figure 4. The insertion of boron into the blanket clearly reduces the tritium production level. After the tritium surplus ramping, the boron concentration oscilation overshoot is in-

9 Figure 4: Boron concentration (appm) within the WCCB and the associated TBR. tially approximately 30%, however over a period of 3 months this attenuates to just 3%. 5. Conclusion Tritium sustainability and sufficiency is an absolute requirement for future DEMO power plants. The emphasis to produce enough tritium to sustain a fusion reactor may result in over production of tritium, which may impact on the risk and storage costs associated with large quantities of surplus tritium. Thus, future DEMO fusion power plants may require a method to manage tritium production. This work has demonstrated that boron injection has the potential to manage tritium production within a water cooled ceramic breeder. A PID controller has been shown to manage the injection of boric acid in to the coolant of a WCCB blanket. Oscillation within the boric acid concentration and TBR have been shown to be minimal and attenuate at an acceptable rate provided the PID control constants are optimised. In the next stage of this work, the feasability of boron injection into a HCPB DEMO will be studied. Boron trifluoride ( 10 BF 3 ) is one of a only a few gases that are suitable as a gaseous neutron poison. Future studies will involve more detailed reactor geometries and will include more sophisticated controllers Acknowledgements This work was supported by United Kingdom Engineering and Sciences Research Council. [1] L. A. El-Guebaly, S. Malang, Toward the ultimate goal of tritium selfsufficiency: Technical issues and requirements imposed on aries advanced power plants, Fusion Engineering and Design 84 (12) (2009) doi: /j.fusengdes [2] W. R. Meier, Tritium breeding management in the high yield lithium injection fusion energy chamber, Nuclear Technology 52 (1981) [3] W. R. Meier, Build-up of tritium in a liquid-lithium breeding blanket for an inertial confinement fusion chamber, 9th Symposium on Engineering Problems of Fusion Research. [4] Mikio Enoeda et al, Development of water cooled ceramic breeder test blanket module in Japan, Fusion Engineering and Design 87 (7-8) (2012) [5] Masato Akiba, Mikio Enoeda, and Satoru Tanaka, Overveiw of the TBM r & d activities in Japan, Fusion Engineering and Design 85 (10-12) (2010) [6] W. M. Stacey, Nuclear Reactor Physics, Wiley, year =. [7] L. Morgan, J. Pasley, The impact of time dependant spectra on fusion blanket burn-up, Fusion Engineering and Design (0) (2013). doi: /j.fusengdes [8] S. Z. A. Aures, L.W. Packer, Tritium self sufficiency of hcpb blanket modules for DEMO considering time-varying neutron flux spectra and materials composition., SOFT. [9] Deuterium-tritium fuel self-sufficiency in fusion reactors, Fusion Technology 9 (1986) [10] S. Sato, T. Nishitani, C. Konno, Effects of lithium burn-up on TBR in DEMO reactor SlimCS, Fusion Engineering and Design 87 (5-6) (2012) , ce:title Tenth International Symposium on Fusion Nuclear Technology (ISFNT-10) /ce:title. doi: /j.fusengdes [11] I. Palermo, J. Gmez-Ros, G. Veredas, J. Sanz, L. Sedano, Neutronic design analyses for a dual-coolant blanket concept: Optimization for a fusion reactor DEMO, Fusion Engineering and Design 87 (7-8) (2012) , ce:title Tenth International Symposium on Fusion Nuclear Technology (ISFNT-10) /ce:title. doi: /j.fusengdes [12] J. Cataln, F. Ogando, J. Sanz, I. Palermo, G. Veredas, J. Gmez-Ros, L. Sedano, Neutronic analysis of a dual He/LiPb coolant breeding blanket for DEMO, Fusion Engineering and Design 86 (2011) doi: /j.fusengdes [13] S. Z. L.W. Packer, R. Pampin, Tritium self-sufficiency time and inventory evolution for solid-type breeding blanket materials for DEMO, Journal of Nuclear Materials 417 (1-3) (2011) [14] X-5 Monte Carlo Team, X-5 Monte Carlo Team: MCNP - A general monte carlo N-Particle Transport Code. [15] EASY-II, European Activation SYstem, [16] K. J. Astrom, Control System Design, [17] Y. C. Dingyu Xue, D. P. Atherton, Control System Design, SIAM, 2007.