Pebble Bed Breeding Blanket Research Articles

Experimental investigation of the corrosion behavior of Eurofer97 steel in contact with Lithium ceramic breeder pebbles under specific Helium Cooled Pebble Bed breeding zone atmosphere

For sub-sized fatigue specimens made of EUROFER97 in unconstrained contact to ceramic breeder pebbles, exposed to purge gas conditions for different durations in an oven, a chemical surface attack was observed which led to a significant reduction of fatigue lifetime [Aktaa et al., 2020]. To better reproduce the flow of the purge gas in the breeding zone of a Helium-Cooled Pebble Bed Breeding Blanket, the experiment was now repeated by placing the same kind of samples in the helium loop HELOKA HEMAT, where the gas is permanently circulated in a closed circuit, and its water content is controllable. The composition of the gas was monitored with a mass spectrometer and humidity sensors. The samples were exposed to a mixture of helium and 0.1vol% hydrogen at DEMO relevant operating conditions (550 °C, 1.2 bars abs.) for a duration of 8, 16, 32, and 64 days, respectively. New sample holders were designed which allow direct contact of the samples with the gas mixture. A special procedure for the handling of the test section was implemented to avoid contact of the hygroscopic pebbles with air humidity during the preparation phase and during the extraction of the samples from the test rig.The present contribution discusses the various considerations on which the test rig design was based, followed by a description of the experimental setup and preliminary results of the testing campaign. These are similar to the results achieved in the oven experiment. In the future, the loop will be used for tests with up to 200 Pa partial pressure of steam in helium; a mixture that is considered relevant for the DEMO purge gas composition.

Thermal-hydraulic study of the EU-DEMO Helium Cooled Pebble Bed Breeding Blanket Primary Heat Transport System

The European DEMO reactor will work under normal operating conditions in accordance with a pulsed duty cycle. However unplanned and planned transients of plasma overpower may occur compromising the integrity of its plasma-facing components structures. Consequently, adopting appropriate tools is essential to accurately and consistently model the thermal-hydraulic response of the involved cooling systems under both normal operating conditions and accidental events. Given this background, the University of Palermo, in collaboration with EUROfusion, started a research work to study the thermal-hydraulic behaviour of the Primary Heat Transport System (PHTS) of the Helium Cooled Pebble Bed Breeding Blanket (HCPB BB) of the DEMO device under steady-state and transient conditions. The activity has been performed adopting a computational approach, employing the thermal-hydraulic system code TRACE version 5.0 patch 6. The key point of the work has been the code-to-code benchmark with the outcomes previously obtained with the RELAP5-3D code, to estimate the impact of the physical models, numerical resolution schemes and modelling techniques adopted on the predictive capabilities of the system codes considered. The models and the analysis results are presented and critically discussed herein.

Advancements in Designing the DEMO Driver Blanket System at the EU DEMO Pre-Conceptual Design Phase: Overview, Challenges and Opportunities

The EU conducted the pre-conceptual design (PCD) phase of the demonstration reactor (DEMO) during 2014–2020 under the framework of the EUROfusion consortium. The current strategy of DEMO design is to bridge the breeding blanket (BB) technology gaps between ITER and a commercial fusion power plant (FPP) by playing the role of a “Component Test Facility” for the BB. Within this strategy, a so-called driver blanket, with nearly full in-vessel surface coverage, will aim at achieving high-level stakeholder requirements of tritium self-sufficiency and power extraction for net electricity production with rather conventional technology and/or operational parameters, while an advanced blanket (or several of them) will aim at demonstrating, with limited coverage, features that are deemed necessary for a commercial FPP. Currently, two driver blanket candidates are being investigated for the EU DEMO, namely the water-cooled lithium lead and the helium-cooled pebble bed breeding blanket concepts. The PCD phase has been characterized not only by the detailed design of the BB systems themselves, but also by their holistic integration in DEMO, prioritizing near-term solutions, in accordance with the idea of a driver blanket. This paper summarizes the status for both BB driver blanket candidates at the end of the PCD phase, including their corresponding tritium extraction and removal (TER) systems, underlining the main achievements and lessons learned, exposing outstanding key system design and R&D challenges and presenting identified opportunities to address those risks during the conceptual design (CD) phase that started in 2021.

The European DEMO Helium Cooled Pebble Bed Breeding Blanket: Design Status at the Conclusion of the Pre-Concept Design Phase

Significant design efforts were undertaken during the Pre-Concept Design (PCD) phase of the European DEMO program to optimize the helium cooled pebble bed (HCPB) breeding blanket. A gate review was conducted for the entire European DEMO program at the conclusion of the PCD phase. This article presents a summary of the design evolution and the rationale behind the HCPB breeding blanket concept for the European DEMO. The main performance metrics, including nuclear, thermal hydraulics, thermal mechanical, and tritium permeation behaviors, are reported. These figures demonstrate that the HCPB breeding blanket is a highly effective tritium-breeding and robust driver blanket concept for the European DEMO. In addition, three alternative concepts of interest were explored. Furthermore, this article outlines the upcoming design and R&D activities for the HCPB breeding blanket during the Concept Design phase (2021–2027).

Modeling of the HCPB Helium Coolant Purification System for EU-DEMO: Process Simulations of Molecular Sieves and NEG Sorbents

The Helium Cooled Pebble Bed breeding blanket of the EU-DEMO foresees continuous processing of a small fraction of the helium coolant in the coolant purification system (CPS) to counteract buildup of tritium and impurities. For this system, two different process variants are currently considered. The first is based on the conversion of all hydrogen species into water using copper oxide beds and the subsequent water adsorption over zeolite molecular sieve (ZMS) beds. The alternative process foresees the direct sorption of hydrogens onto novel ZAO® non-evaporable getter (NEG) materials. The ZMS beds and the NEG beds are operated batchwise, but alternating schemes with an absorption (operation) phase and a desorption (regeneration) phase result in a pseudocontinuous process. Transient process simulations have been developed to evaluate the performance and impact of the different variants on downstream systems in the fuel cycle. In this contribution, these process models for the preconceptual design of both variants are presented and evaluated. For the reference designs proposed for each system, they have been found to satisfy the requirements of achieving 90% efficiency. This modeling then lays the foundation for optimization of the conventional process and outlines further research demand regarding the application of NEG materials needed to progress toward the concept design of the CPS process.

Shielding Design Optimization of the Helium-Cooled Pebble Bed Breeding Blanket for the EU DEMO Fusion Reactor

The helium-cooled pebble bed (HCPB) breeding blanket (BB) is one of the two driver-blanket candidates for the European DEMO fusion reactor. Recent design activities were focused, among other objectives, on the achievement of an efficient shielding system to adequately protect the vacuum vessel (VV) and toroidal field coils (TFCs). Several shielding options have been studied in terms of architecture (e.g., in-BB shield and ex-BB shield) and materials (e.g., B4C, WC, WB, YHx, and ZrHx). In this study, the B4C material was selected as the most attractive option considering not only shielding performance but also availability, industrialization, experience, and cost factors. Subsequently, we performed a parametric study by implementing different thicknesses of a B4C external shield and reporting information of its effect on shielding performance, structural behavior, swelling and tritium breeding. Furthermore, a detailed structure for the VV was developed considering an internal layered configuration comprising steels/water with different boron contents. Corresponding shielding analyses were conducted regarding influence on neutron attenuation when implementing such a VV structure for both the baseline consolidated design of the HCPB and one of the previously developed and improved BSS configurations. The most critical responses (neutron flux and dpa) were fully established only using 10 cm B4C and an improved VV configuration.

Hydrogen permeation from F82H wall of ceramic breeder pebble bed: The effect of surface corrosion

Understanding the permeation behavior of tritium from a pebble bed breeding blanket is essential for establishing a self-sufficient fuel cycle in a nuclear fusion reactor. It is known that double corrosion layers forms on reduced activation ferritic-martensitic (RAFM) steel surface by gas release from a ceramic breeder material, however, its effect on hydrogen permeation behavior has not been elucidated. Herein, in-situ measurement of hydrogen permeation through an F82H RAFM wall of a ceramic breeder pebble bed was performed under H2-added sweep gas conditions. The corrosion layer formed on the F82H sample had a dense microstructure, which reduced hydrogen permeation flux at least by one order of magnitude. The permeation reduction factors were 20–50 at the water-coolant temperature of a blanket. A self-repairing ability is expected for the surface oxide layer as the corrosion occurs spontaneously inside a breeding blanket.

Model Development and Transient Analysis of the HCPB BB BOP DEMO Configuration Using the Apros System Code

Extensive modeling and analytical work has been carried out considering the Helium-Cooled Pebble Bed Breeding Blanket (HCPB BB) Balance Of Plant (BOP) configuration of the Demonstration Power Plant (DEMO) using the Apros system code, developed by VTT Technical Research Centre of Finland Ltd. and Fortum. The integral plant model of the HCPB BB plant has been improved with respect to the blanket and steam generator models. Based on HCPB-BL2017 v1 data, reported in 2019, the blanket has been remodeled by separate Apros process components, dedicated to average inboard and outboard segments, where the power deposition scheme of the breeding units took into account the output of high-fidelity neutronic analyses. A new helical coil steam generator model has been developed for primary–secondary system coupling using CAD data provided by EUROfusion partner University of Palermo. Transient analyses have been performed with Apros on the plant configuration that utilizes a molten salt technology-based small Energy Storage System (ESS).

Stress Analysis of HCPB BB under Ex-Vessel LOCA Condition

Helium Cooled Pebble Bed Breeding Blanket (HCPB BB) is a kind of concept for the European demonstration fusion reactor (DEMO). The blanket attachment system plays an important role in the mechanical connection of the BB and vacuum vessel. Typically, the mechanical and thermal loads should meet the requirement to avoid collapse of the system with off-normal conditions, e.g., under ex-vessel Loss of Coolant Accident (LOCA. This paper investigates the loading requirement corresponding to the maximum stress that can sustain to avoid the LOCA condition. Firstly, a model of the BB is constructed using SolidWorks. Then, stress analysis is carried out based on the cross section of the blanket. Through simulation, the critical condition for the LOCA case and the maximum stress value for the model are obtained. According to the relevant size dimension from the reference, the blanket’s cross section is drawn, and one can get the stress field under the ex-vessel LOCA through stress analysis. The stress distribution under the ex-vessel LOCA condition is simulated to find out the maximum stress field that the blanket can sustain through this paper. The significance is to predict the possible conditions leading to an accident and find possible methods to avoid them.

Heterogeneous design and mechanical analysis of HELIAS 5‐B helium‐cooled pebble bed breeding blanket concept

One of the most challenging objectives of the European research concerning nuclear fusion technology, promoted by the EUROfusion consortium, is to bring stellarator-type nuclear fusion devices to maturity. To this purpose, studies on a large HELIcal-axis advanced stellarator (HELIAS), extrapolated from Wendelstein 7-X and based on a 5-fold symmetry (HELIAS 5-B), are currently ongoing. The HELIAS 5-B stellarator reactor will be endowed with a breeding blanket (BB) system to allow for the self-sustainability of the nuclear fusion reaction and make it suitable for electricity generation. In this paper, we present the first ever heterogeneous mechanical design and the preliminary structural assessment of a bean-shaped ring of a HELIAS 5-B BB sector. The proposed mechanical design, which is based on the helium-cooled pebble bed (HCPB) BB concept and developed according to the “sandwich” architecture, foresees an actively cooled segment box connected to a back-supporting structure equipped with manifolds. The internal region (breeding zone) is reinforced by actively cooled steel plates. The proposed heterogeneous design was checked against nominal loads and an in-box loss of coolant accidental scenario, which is a typical design driver for BBs. The assessment has been performed according to the RCC-MRx structural design code. Our results are herewith presented and critically discussed, focusing on the potential follow-up of the HELIAS 5-B HCPB BB design.

Preliminary Accident Analysis of Ex-Vessel LOCA for the European DEMO HCPB Blanket Concept

For the helium-cooled pebble bed breeding blanket concept improved in 2016 and the associated primary heat transfer system (PHTS) following EU DEMO Baseline 2015, an ex-vessel loss-of-coolant accident (LOCA) has been investigated with the assumption of a double-ended guillotine break of a main pipe in an outboard (OB) loop of the PHTS. The break leads to helium blowdown into the tokamak cooling room. A fast plasma shutdown followed by a plasma disruption is activated after the detection of the LOCA due to the design basis accident. Regarding three affected first-wall (FW) areas in one or two OB loops, three main cases are considered. If the FW temperature reaches the defined temperature limit of 1000°C, the FW is assumed to be failed such that an in-vessel LOCA results. In total five scenarios are simulated using MELCOR 1.8.6 for fusion with respect to the affected FW areas, mitigated or unmitigated plasma disruption conditions, the options of the dry or wet suppression tank, and the transport of source terms performed in the case of the beyond design basis accident without the plasma shutdown. The transient results are discussed for the time evolution of the accident sequences, pressurization in the systems, temperature behavior in volumes and structures, and tritium and dust transport behavior.

Pre-conceptual design of EU-DEMO divertor primary heat transfer systems

In the frame of the activities promoted and encouraged by the EUROfusion Power Plant Physics and Technology (PPPT) department aimed at developing the EU-DEMO fusion reactor, strong emphasis has been recently addressed to the whole Balance of Plant (BoP) which represents the set of systems devoted to convert the plasma generated thermal power into electricity and to deliver it to the grid. Among these systems, the Divertor Primary Heat Transfer Systems (PHTSs) are intended to feed coolant to the two main components of the Divertor assembly, namely the Plasma Facing Components (PFCs) and the Cassette Body (CB). Since the DEMO Divertor must withstand high heat flux loads together with a considerable neutron deposited power, very tight tolerances may be allowed to the coolant inlet conditions. Therefore, the design of reliable PHTSs is of the utmost importance towards the development of an EU-DEMO fusion reactor. Within this framework, a study has been jointly carried out by University of Palermo, Ansaldo Nucleare and CREATE to design the Ex-Vessel PHTSs of both the PFCs and the CB for a DEMO reactor equipped with a Helium Cooled Pebble Bed Breeding Blanket. The paper describes criteria and rationale followed with the aim to achieve simple PHTS designs based on the adoption of easy-to-manufacture main components avoiding too much extrapolation from the state-of-art technology. Results of preliminary thermal-hydraulic calculations carried out to size heat exchangers, pressurizers, piping and pumps are presented and critically discussed, with particular attention to those integration, safety and feasibility constraints that may deeply affect the design of such components. Finally, the evaluation of PHTS key parameters as total pumping power and coolant inventory is reported.

Benchmark of Serpent-2 with MCNP: Application to European DEMO HCPB breeding blanket

Abstract Magnetic confinement fusion reactors such as the European fusion demonstration reactor (EU DEMO) have a very complex geometry, which makes the modelling of the nuclear analyses rather tedious and time consuming. The Monte Carlo neutron transport code Serpent-2 developed by VTT in Finland, is able to directly import CAD geometry, then is considered an attractive alternative to the Monte Carlo N-Particle Transport code (MCNP) for its application on fusion reactors. In the first step, a comparison activity is carried out to benchmark the Serpent-2 with MCNP calculations. Both Serpent-2 and MCNP are used to perform the 3D Monte Carlo neutron and photon coupled transport making use of the EU-DEMO BL2017 model and a grid discrete volume source. A very useful conversion tool (called csg2csg) developed by UKAEA researchers was used to translate the MCNP input model to the Serpent-2 input model. The nuclear performance of the EU DEMO Helium Cooled Pebble Bed breeding blanket, such as tritium breeding ratio (TBR), neutron and photon flux spectra, and nuclear heating are analyzed, and the corresponding results are compared between Serpent-2 and MCNP. This study shows that Serpent-2 is suitable for the application of nuclear analysis for the EU DEMO HCPB breeding blanket.

Advancements in the Helium-Cooled Pebble Bed Breeding Blanket for the EU DEMO: Holistic Design Approach and Lessons Learned

The helium-cooled pebble bed (HCPB) blanket is one of the two concepts proposed as a driver blanket for the European Union Demonstration Fusion Power Reactor (EU DEMO). In contrast to past conceptual design studies, in the frame of the current Power Plant Physics and Technology of the EUROfusion Consortium, the ongoing EU DEMO preconceptual design activities have adopted a holistic and integrated (i.e., systems engineering) design approach. As a consequence of this new approach, many interfaces and requirements have been identified, some of them driving the design of the blankets. This paper shows the advancements in the HCPB breeding blanket and describes the lessons learned after implementing the new approach. This new set of requirements has led us to reconsider fundamental aspects of the HCPB blanket design, especially in the way of how the heat is extracted from the blanket. Among others, the requirement to achieve a mature balance of plant (BOP) system plays a central role as a key design driver and has forced us to reduce pressure drops in the breeding blanket. In this regard, the blanket has been redesigned, leading to an enhanced concept based on single-module segments with a hexagonal matrix of fuel-breeder pins. Both the fuel-breeder pins and the first wall (FW) are equipped with turbulence promoters (augmented wall roughness in the fuel-breeder pins and V-ribs in the FW), following a similar idea as in the past MAGNOX, Advanced Gas Reactor (AGR), and Gas Cooled Reactor (GCR) programs in fission. This has led to minimizing the pressure drops while maximizing the heat transfer. Also, the blanket outlet temperature has been extended to 520°C, following the same principle as in Generation IV’s GCRs of maximizing the temperature difference across the core to minimize the reactor mass flow rate and thus the circulating power. All these features have led to a remarkably low plant circulating power (80 to 90 MW) and the required power per helium blower (5 to 6 MW), which potentially solves the key long-standing problem of the BOP technology readiness level for an ≈2.4-GW(thermal) helium-cooled DEMO reactor.

Analysis of the Flow Distribution in the Back Supporting Structure Manifolds of the HCPB Breeding Blanket for the EU DEMO Fusion Reactor

The European Union Demonstration Fusion Power Reactor (EU DEMO) is facing its preconceptual design phase. In this phase, the research and development activities make extensive use of computational tools, to, e.g., verify the design calculations or to perform parametric analyses aimed at optimization. The design of the breeding blanket (BB), which will be a first-of-a-kind component in EU DEMO, is supported from the thermal-hydraulic point of view by local three-dimensional (3-D) computational fluid dynamics (CFD) analyses, mainly aimed at verifying the heat removal capabilities of the system, and by analyses at the system level using one-dimensional (1-D) codes.This work presents the development and application of a detailed 1-D model of the coolant manifolds for the helium-cooled pebble bed BB concept for EU DEMO. This model, implemented in the GEneral Tokamak THErmal-hydraulic Model (GETTHEM), allows fast analyses to be performed at the global level but still maintain a good level of detail concerning the coolant distribution. The first results obtained with the model prove that 3-D CFD analyses of the manifolds may provide misleading results due to nonrepresentative boundary conditions (BCs), which must be used to avoid having a domain that is too complex. The application of a global model, which is indeed characterized exploiting local analyses, can in turn provide better BCs to the detailed 3-D CFD analyses.

Preliminary CAD implementation of EU-DEMO primary heat transfer systems for HCPB breeding blanket option

This paper focuses on the 3D CAD implementation of the pipework and the main equipment of the Primary Heat Transfer System of EU-DEMO fusion power plant. In particular, the systems related to the Helium-Cooled Pebble Bed Breeding Blanket option are considered here. During the pulse operation, the breeding blanket modules will be the main thermal power source; Divertor and the Vacuum Vessel will contribute in the definition of the total reactor power. All the In-Vessel generated power is rejected to the Power Conversion System through a molten salt Intermediate Heat Transport System. The latter is equipped with an Energy Storage System to allow for continuous operation also during the dwell time. The layout of the pipework considered the major design constraints, such as the reinforced concrete structures and the dimensions of the main pieces of equipment. The size of the pipes was selected based on operating pressure, temperature and allowable velocity. Results provide a first estimation of length and weight of the piping systems as well as their space occupancy. These data are useful to evaluate the coolant inventory, as well as the expected pressure drops and plant cost.

Thermal-hydraulic analysis of the First Wall of a CO2 cooled pebble bed breeding blanket for the EU-DEMO

The Demonstration Fusion Reactor in Europe (EU-DEMO) is considered to be the last step before building a commercial fusion power plant. Up to now, He has been chosen as gas coolant for three of the four candidate breeding blankets proposed for the EU-DEMO, being the Helium Cooled Pebble Bed (HCPB) among them. The choice of He coolant is supported, among other features, by its superior heat transfer capabilities, its transparency to neutrons and chemical inertness. However, the use of this coolant in DEMO poses some feasibility questions, like the large coolant inventory, resulting in possible safety issues, as well as the large circulating power. On the other side, CO2 has been the fluid coolant choice in the nuclear fission industry for gas-cooled reactors since the 1950s. The key advantage of this gas is its larger density, which results in more compact components, lower coolant inventories, lower circulating powers and thus better plant thermal efficiency. This, coupled also with a relatively high molecular stability at moderate temperatures (up to 600 °C), its larger availability and lower price make CO2 an attractive coolant for a pebble bed breeding blanket, as an alternative to He.As a Plasma Facing Component, the First wall (FW) is one of the components under the most challenging environment. Currently, the FW region at the upper port inboard location is expected to suffer the most critical heat flux loads. Indeed, a peak charged particle heat flux 0.99 MW/m2 together with a radiation heat flux of 0.18 MW/m2 is foreseen at this location. This paper reports the thermal-hydraulic analyses carried out to find the practical heat flux limits by using CO2 as coolant. The analyses have been performed by using the commercial computational fluid dynamics (CFD) code ANSYS CFX. Different augmented technical wall roughness areas have been considered, for a best heat transfer-to-pressure drop ratio. Sensitivity analyses have been performed as well for higher peak particle heat fluxes, evidencing that a CO2-cooled FW can dissipate peak heat fluxes up to 1.8 MW/m2 keeping the circulator power in a reasonable range.

Comparative analysis of the efficiency of a CO2-cooled and a He-cooled pebble bed breeding blanket for the EU DEMO fusion reactor

The EU DEMO is considered to be the last step before building a commercial fusion power plant. Up to now the Helium Cooled Pebble Bed (HCPB) Breeding Blanket (BB) is one of the four BB concepts being investigated in the EU for their possible implementation in the EU DEMO.While He is the reference gas cooling for the BBs in DEMO, CO2 has been the coolant choice in the nuclear fission industry for gas-cooled reactors during the past 60 years. In the recent years, it is also gaining interest as a secondary coolant for the Balance of Plant (BoP) to replace the traditional vapor (Rankine) cycle. Despite the lower heat capacity and thermal conductivity of CO2 with respect to He, the larger density of the former can lead to an advantageous performance on heat dissipation rate per unit of pumping power. This feature, coupled with its relative transparency to neutrons, the molecular stability at the operating temperatures of the BB and the industrial scale ready turbomachinery makes CO2 an attractive alternative to He.The aim of this paper is to make a parametric study of the gross and net efficiencies and the design of a simplified BoP of a DEMO power plant based on a CO2-cooled compared to the reference EU DEMO HCPB BB, as a function of the BB outlet temperature and turbine inlet pressure. From this study it is concluded that the vacuum vessel and the divertor heat sources are benefitial in order to improve the net plant efficiency for both coolants. Also, it has been observed that the plant net efficiency is better for the case of a PHTS with CO2 as heat transfer fluid under the assumption that both BBs share the same geometry. In the case that the BBs do not share the same geometry, the study concludes that a PHTS with CO2 leaves a pressure drop margin of about 1.5 before reaching the same net efficiency as a HCPB BoP DEMO plant.

Thermal-Hydraulic Analysis of the EU DEMO Helium-Cooled Pebble Bed Breeding Blanket Using the GETTHEM Code

The general tokamak thermal-hydraulic model (GETTHEM) has been updated to the most recent version of the EU DEMO helium-cooled pebble bed breeding blanket (BB) design. The GETTHEM results are first benchmarked in a controlled case against the results of 3-D computational fluid-dynamics computations, showing an acceptable accuracy despite the inherent simplifications in the GETTHEM model. GETTHEM is then applied to the evaluation of the poloidal hot spot temperature distribution in an entire BB segment, showing that the maximum temperature in the EUROFER structures overcomes the design limit of 550 °C by more than 50 °C in some blanket modules. A possible mitigation strategy is then proposed and analyzed, based on the idea of cooling the first wall in parallel with the breeding zone, showing that this solution would allow having the EUROFER in its working temperature range in the entire segment, although at the expense of a larger pressure drop.

Comparison of MFI-ZSM5 and NaA zeolite-type tubular membranes for the separation of water vapour from helium for tritium processes in future fusion reactors

Zeolite membranes have been increasingly used for gas separation applications, where the most studied are the MFI and NaA types. Typically, NaA membranes are used for dehydration applications whereas MFI are used for the separation of large/small molecules (e.g., CO2/H2). In the Tritium Laboratory Karlsruhe (TLK), both NaA and MFI-ZSM5 were selected for tritium processes where both molecular and oxidized tritiated species diluted in helium have to be processed in a continuous mode, as it is the case of the Tritium Extraction System of the Helium Cooled Pebble Bed breeding blanket. This paper presents and discusses the results obtained for the separation of a binary mixture of water vapour and helium, down to 0.2vol%, where hydrogen replaces its radioactive counterpart. These experiments were performed in the temperature range between 25 and 100°C. NaA demonstrates much better performances for feed concentrations ≥5vol%, eventually reaching separation factors higher by a factor of 100 than that of MFI-ZSM5. However, the MFI-ZSM5 membrane must be selected instead when moisture levels are ≤1vol%. In fact, at these conditions, the H2O/He separation factor for NaA barely exceeds 2, whereas it can be as high as 9.6 for MFI-ZSM5. Moreover, rather high permeances for H2O of around 1–2μmolm−1s−2Pa−1 were measured for both membranes. In the case of helium, the permeances were highly dependent on the temperature for high water contents, reaching values as low as 1×10−3μmolm−1s−2Pa−1 for the NaA membrane.