Tritium management in ITER test blanket systems port cell for maintenance operations

Four Test Blanket Systems (TBS) will be tested in the International Thermonuclear Experimental Reactor equatorial ports #16 and #18 to verify tritium breeding and heat extraction technology. A significant quantity of tritium would be produced in TBM, and partly released into the port cell from the pipework of TBS or other high-temperature components due to its strong mobility and high permeation. The port cell should be accessible during equipment maintenance and human intervention. This work built a multi-dimensional geometric model to characterize HTO transport in the port cell, absorption/desorption, and diffusion in walls and discussed the effect of paint thickness, ventilation rate, source term, and epoxy properties on detritiation efficiency. The results suggest that a 0.1–0.16 mm paint with the lowest HTO solubility is optimal from the compromise between quick cleanup and tritiated waste decommission. A higher ventilation rate could accelerate detritiation while minimizing the radioactive source by a tritium-resisting layer is the most direct method. The optimized design options for reducing the time required to reach 1 DAC in 12 h still need further discussion because of the delayed HTO source from epoxy paint and dead zone of the flow field.


Introduction
ITER will be the first fusion facility to operate a long pulse D-T burning plasma and breed tritium fuel [1,2].TBM will be installed in the equatorial ports, and their ancillary systems will be in the ancillary equipment unit (AEU) [3].The Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.port cell is classified as C3 confinement class, where the occupational radiation exposure for these maintenance operations should be designed to be as low as reasonably achievable.TBM would produce a significant quantity of tritium, and partly released into the port cell from the pipework or other high-temperature components due to the strong mobility and high permeation.The Detritiation System (DS) should be able to reduce the tritium concentration below the admissible limit for equipment maintenance and human intervention in the port cell (regulatory maximum allowable value is 1 DAC (Derived Air Concentration), for tritium 1 DAC corresponds to 3.4 × 10 5 Bq m −3 of air).It should be reached in less than 12 h after the plasma shutdown.Tritium management in ITER is one of the main safety engineering issues [4][5][6].Thus, several researches have been conducted on the tritium inventory and transport in TBM port cells.
In 2014, Giancarli et al delineated an analytical approach to assess tritium concentration variation in port cell after the plasma shutdown [7].A good compromise to reach the target that tritium concentration in air reduce to 1 DAC in less than 12 h is given by the case with 40 m 3 h −1 at normal DS and 60 m 3 h −1 for enhanced DS.They also tried a total transient 3D model to evaluate the detritiation process, which was impeded by the enormous computational resource.Chen et al estimated the tritium release in some major components of China Helium-cooled Ceramic Breeder (HCCB) Test Blanket Systems (TBS) through permeation and natural leakage models from 2015 [8,9].The results showed that due to the tritium permeation barrier coating for tritium confinement in some tritium containments, the total tritium release to the environment by permeation would be kept well below 2 Ci/full-power day.Merril et al used TMAP 4 to assess the tritium contamination evolution after 12 d of continuous pulsed plasma in the port cell #16 in 2016 [10].They proposed that it needs an encapsulation for all components to limit the tritium release for human access in less than 12 h.Brahmbhatt et al built a steadystate diffusion-limited permeation model in 2017 that neglects surface effects and has been used to study the tritium transport from Lithium Lead Ceramic Breeder (LLCB) TBS and its ancillary systems into process rooms [11].The tritium distribution in Vacuum Vessel, Pipe Forest Area, Port Cell, Pipe Chase Area, and Tritium Process Room in the L-2 level has been estimated, and 0.0324 mg T d −1 would be released into the port cell.Ricapito et al built an advanced tritium transport model for Helium Cooled Pebble Bed (HCPB) and Water Cooled Lithium Lead (WCLL) TBS based on the EcosimPro platform in 2018, which implements a 1D dynamic model at the full TBS level, with a tritium transport regime that considers the surface effects of atomic dissociation and molecular recombination [12].This code could predict the dynamic inventory and permeation through the TBS components after full validation through parametric analyses and benchmarking against experimental results [10].
All the tritium management is based on 0/1 D models or analytical methods [7][8][9][10][11][12], which cannot obtain the accurate distribution of tritium if there is ample space with a complex flow field.Besides, the tritium from TBS makes an HTO environment in the port cell plant.Some HTO would be adsorbed/absorbed on/in porous surfaces and migrate inside concrete due to the soaking effect [13][14][15][16][17]. The tritium concentration in the port cell should be reduced to the permissible limit for human access during maintenance operations.However, this assessment is challenged by multi-dimensional geometry and multi-physics environments.Therefore, this paper presents a 3D + 1D geometric model to characterize the HTO behavior in port cell air, absorption/desorption, and diffusion in the wall based on the finite element method.The analysis has been performed for equatorial port #16 because it features the highest tritium permeation due to low-velocity LiPb pipes [18].Additionally, the effect of epoxy paint, ventilation rate, and the inventory in long-term operation have been discussed to provide data support in the engineering design of the TBM port cell plant.

Methodology
This study focuses on the effect of paint and concrete on the time required to reach 1 DAC after shutdown (for maintenance access), where HTO transport between paint and air is essential.Since the paint on wall and port cell air have vastly different dimensions, a total 3D model would consume too many computing resources to characterize the transition at the interface accurately.Therefore, a 3D + 1D geometric model was developed to represent the HTO transport in TBM port cell, as shown in figure 1.The tritium transport is mainly influenced by air convection and adsorption/desorption on the porous surface.So, a detailed 3D model of the port cell is necessary for reproducing the actual flow field.The flow field reaches a steady state for a while once the DS functions.Then, the tritium transport was weakly coupled to the flow field by equation ( 1), where the typical mass conversation model considers HTO diffusion, convection, decay in air, and the leakage source from TBS [19], where, D i : Diffusivity of species i (m 2 s −1 ), c i : hydrogen isotope i concentration in free (mol * m −3 ), ⃗ u: velocity of fluid, R: hydrogen isotope source term, λ: decay factor.The HTO in paint and concrete is only transported by diffusion, which could be described as a 1D model because it is much thinner in thickness than others.Equation (1) could also explain the diffusion process in the wall when the convection is eliminated.The hydrogen chemical potential µ continues at the paint and concrete interface, defined as (2) [20][21][22], where, µ 0 represents the hydrogen standard chemical potential and S is hydrogen solubility in the considered material.
A simple dynamic model (3) described the HTO transport between air and epoxy paint, automatically managing the absorption and desorption process [23], where flux[m 2 s −1 ] is HTO transport flux at the interface, ν[m s −1 ] is the dynamic factor (0.1 (m s −1 ) in simulation), c paint and c Air are the HTO concentration at the surface of paint and air, respectively.Temperature is also an essential factor affecting the transport of HTO in both air and walls [24,25].The predicted heat source of 11 kW is lower than the Local Air Cooling (LAC) design value of 43 kW [7].Thus, the thermal distribution in the TBM port cell remains stable and fluctuates around room

Steady-state flow field
The LAC system works only if heat is released into the plant [4].Thus, only the DS effect to flow field is considered in this study.A steady-state flow field would be established after the DS functioned for a while.The detritiation outlet boundary condition was set to 40 m 3 h −1 , while the inlet was set to −100 Pa to ensure a negative pressure environment in the plant [7]. Figure 2 illustrates the steady-state flow field in the TBM port cell.The maximum velocity is near the inlet surface and is directly related to the shape and size of the entrance.The flow is relatively weak in the Port Interface area, with turbulence occupying most of the space.The 2D plane flow reveals several dead zones (flow immobile) distributed in the corners or near the wall surfaces.

Delay source in detritiation
The porosity of epoxy paint and concrete allows for HTO absorption and desorption, resulting in a substantial HTO inventory over the prolonged operation.Consequently, the HTO in the wall represents a potential tritium source during short-term maintenance (STM) and may produce a quantity of radioactive concrete during decommissioning.To quantitatively evaluate the effect of paint and concrete on the HTO inventory, a series of controlled trials were conducted in a continuous back-to-back pulse operation cycle.The ventilation rate in the port cell remained at 40 m3 h −1 during both normal operation and STM conditions.The HTO released from the connecting pipes between PF and AEU was 0.669 mg T per full power day [10].The ventilation rate in the port cell remained at 40 m3 h −1 during both normal operation and STM  conditions.Besides, to display the HTO variation until the accessible limit, the STM period is extended to 12 d, whereas the actual plan is 3-4 d.
Figure 3 shows the average HTO concentration variation in port cell#16 with and without epoxy paint (0.16 mm Holland Paint) and concrete (50 mm) under consideration.The results indicate that there is only a slight difference (<1%) at equilibrium during pulsed plasma operation, although the time to reach equilibrium is prolonged due to the transport of HTO in paint and concrete in the first four days.However, a significant discrepancy is observed after plasma shutdown.When the effect of paint and concrete is ignored, a large amount of HTO is rapidly removed by fresh air, and it only takes 1.36 d to reach 1 DAC for human access.This analysis takes the inhibiting effect of dead zones on detritiation efficiency into account, resulting in a 0.82 d delay compared to other analytical methods that do ignore this effect [7].When the impact of paint and concrete is considered, reaching 1 DAC is extended to 14.09 d because building materials become a delayed source and continuously release HTO.Thus, it is necessary to count the effect of building materials during tritium safety evaluation in port cell.

Effect of paint thickness
Although the solubility of HTO in Holland paint and concrete is the same, it diffuses much faster in concrete (almost 1000 times faster) than in paint.It means that HTO is more easily absorbed in paint than in concrete.Therefore, the thickness of the paint coating is an essential factor in determining HTO performance in building materials.To evaluate the effect of paint thickness, a ventilation rate of 40 m 3 h −1 and a tritium source of 0.28 mg T d −1 were used [18], and the different paint thicknesses are shown in table 2.
Figures 4(a) and (b) show the concentration of HTO in air and the quantity of HTO in paint/concrete, respectively.The HTO in the air is almost at the same concentration, while the inventory of HTO in the paint and concrete varies for each test during the 12 d of back-to-back plasma.The HTO inventory in the paint will reach saturation if the paint is not thick enough.For example, the paint with a thickness of 0.05 mm and 0.1 mm took 2.5 d and 7 d, respectively, to reach HTO saturation, as shown in figure 4.Moreover, the thinner the paint, the more HTO stored in the concrete, which is challenging to eliminate and can lead to the production of tritiated construction during decommissioning.The HTO in paint will be released into the air and concrete when the plasma shuts down.The thicker the paint is, the more HTO is stored in it and the longer it takes to reach the admissible limit for human access, as the paint becomes a delayed source.Therefore, in order to ensure quick entrance after shutdown and less tritiated construction during decommissioning, the thickness of paint is better in the range of 0.1 mm-0.16mm.

Effect of paint properties
HTO released from the TBS would interact with the epoxy paint, diffuse, and be stored to become a delayed tritium source during maintenance.ITER has proposed several epoxy paints with discrepancies in the tritium solubility and diffusion coefficient, as listed in table 3. The effect of epoxy paints on tritium management is evaluated at a 40 m 3 h −1 ventilation rate, 0.28 mg T d −1 , and 0.16 mm thickness.
Figure 5(a) presents the HTO concentration in port cell air for the six paints during 12 d of back-to-back plasma and 12 d of shutdown.While the tritium concentrations are at the same level during plasma operation, there is a discrepancy once the plasma shuts down.The detritiation time is negatively related to the HTO solubility, as evidenced by the TFTR paint requiring 6.39 d to reach 1 DAC, while the Holland paint needs other more 4.51 d.It illustrates that paint properties also play a role in determining the detritiation time.Moreover, the HTO inventory in both the paint and concrete is predominantly influenced by HTO solubility.Figure 5(b) demonstrates the HTO inventory, which is almost linearly related to solubility.As more tritium accumulates in the paint during pulsed plasma, more tritium is released into the port cell air, and it takes longer to reach an acceptable limit at maintenance.These results further confirm that tritium in the paint can act as a delayed source during maintenance.Therefore, paints with lower HTO solubility are the optimum choice from a tritium safety perspective.

Effect of ventilation rate
Ventilation rate is also a significant factor in reducing the detritiation time for human access after plasma shutdown.The higher the ventilation rate, the more efficiently HTO in the  air can be discharged from the plant, and the adsorbed HTO can be released due to the concentration gradient.A Holland epoxy paint with a thickness of 0.16 mm and a tritium source of 0.669 mg T d −1 was used to evaluate the effect of the ventilation rate.Figures 6(a) and (b) show the concentrations of HTO in air and the quantity of HTO in paint/concrete, respectively.The ventilation rate significantly impacts the HTO inventory, which linearly declines with the increase of ventilation rate.The equilibrium HTO concentration in the air during the 12 day pulsed plasma decreases from 770.9 DAC to 329.8 DAC as the ventilation rate increases from 40 m 3 h −1 to 100 m 3 h −1 .The specific results are summarized in table 4. The HTO in the paint and concrete follows the same trend.Once the plasma shuts down, the HTO in the paint divides into air and concrete.When the ventilation rate is significant, the HTO in the concrete has a smaller proportion, which proves that increasing the ventilation rate not only accelerates the ejection of HTO in the air but also reduces tritiated waste.Unfortunately, increasing the ventilation rate also cannot meet the requirement of reducing the HTO concentration to 1 DAC for 12 h, even though it can effectively remove tritium.

Effect of the source term
The most efficient method to reduce tritium inventory in the port cell is to limit the source released from TBS.Three typical possible released sources from tritium safety analysis were chosen to evaluate the detritiation efficiency.0.16 mm Holland epoxy paint and 40 m 3 h −1 ventilation rate were used, and the sources are 0.104 [27], 0.28 [18], 0.669 [10] mg T d −1 , respectively.Figures 7(a) and (b) show the HTO inventory in the air and wall.The HTO concentration in air stabilizes at 120, 323, and 773 DAC during the pulsed plasma, and the HTO in paint accumulates to 0.336, 0.981, and 2.20 mg after 12 d of pulsed plasma, respectively.These two results are linearly related to the released source, proving that minimizing leakage is the most effective way to limit the HTO inventory.The HTO in the paint would be released into the plant or diffused into the concrete after the plasma shutdown.It hinders the detritiation process for safety access.Thus, a tritium-resisting layer was strongly suggested to prevent tube leakage for quick entrance and less radioactive construction waste.

Inventory of HTO in long-term operation
'Can tritium be produced in the blanket and extracted at a rate equal to tritium consumption in the plasma plus losses by radioactive decay from tritium inventories in reactor components?' [28] would be answered by the TBM Program during continuous back-to-back pulsed plasma.Therefore, the inventory of HTO during long-term operation is a significant factor from a tritium safety perspective.The HTO concentration in the air is periodic and has an equilibrium concentration of 291 DAC during plasma operation, which drops to 2.31 DAC after plasma shuts down for 4 d.The HTO in the paint also shows a periodic tendency, with the inventory increasing slightly.It accumulates to 0.287 mg after 480 d and arrives in a steady state.The concrete thickness is 0.5 m, much larger than the paint's.Therefore, the concentration in the concrete is always low, and HTO only transports into the concrete from the paint, whether during operation or maintenance.It accumulates to 6.34 mg after 30 cycles, much larger than the inventory in the paint.The projected lifetime for ITER is 20 years.Then 0.1 g HTO would accumulate in the concrete.

Conclusion
Due to high mobility and significant radioactivity, the tritium management in ITER is one of the main safety challenges.This paper built a multi-dimensional geometric model to characterize HTO transport in port cell #16, and absorption/desorption/diffusion in walls.The main conclusion can be made as follows.
• During operation, the HTO adsorbed by paint would permeate into concrete to produce large tritiated waste.• During STM/LTM, the HTO adsorbed by paints will act as a continuous release source, which makes it challenging to limit the T-concentration below 1 DAC in 12 h.• The dead zones in the corner or near the wall surface will also eliminate the HTO detritiation efficiency in the PC.
From the simulation results and discussion, some suggestions are proposed from the perspective of tritium safety as follows: • Since the hydrogen property of paints strongly dominates the tritium retention and release, the paints with low HTO solubility should be recommended to control the HTO release from paint and inventory in concrete.• For epoxy paint, a thickness of 0.1-0.16mm is suggested by compromising between quick cleanup and tritiated waste decommission.• If possible, metal or metal oxide coatings (such as aluminum/alumina or stainless steel) are a better choice as a concrete wall barrier than epoxy paints from the perspective of tritium management in PC.
Finally, we should also pay attention to the low concentration but large amount of tritiated waste.For ITER or ITER/DEMO scaled DT fusion device, kilograms of tritium operates and cycles in high temperature environment, once HTO/HT contaminates the building, it would be a chronic and intractable problem for decommission, just like tritium water problem after Fukushima accident.

Figure 2 .
Figure 2. The flow field distribution in 3D port cell (a) and 3.56 m plane of height (b).

Figure 3 .
Figure 3.Effect of paint and concrete on the time required to reach 1 DAC after plasma shutdown.

Figure 4 .
Figure 4. Effect of paint thickness on HTO behavior in port cell#16.(a) Shows the HTO concentration, and (b) shows the quantity of HTO in paint and concrete.

Figure 5 .
Figure 5.Effect of paint properties on HTO behavior in port cell#16.(a) Shows the HTO concentration, and (b) shows the quantity of HTO in paint and concrete.

Figure 6 .
Figure 6.Effect of ventilation rate on HTO behavior in port cell#16.(a) Shows the HTO concentration, and (b) shows the quantity of HTO in paint and concrete.

Figure 7 .
Figure 7. Effect of source term on HTO behavior in port cell#16.(a) Shows the HTO concentration, and (b) shows the quantity of HTO in paint and concrete.

Figure 8 .
Figure 8. Inventory of HTO in long-term operation.

Table 2 .
Related variable in the simulation.

Table 3 .
Epoxy paint properties at room temperature.

Table 4 .
The HTO concentration in air and paint after 12 d of pulsed plasma.