Table of contents

This special issue of Nuclear Fusion contains 13 informative papers that were initially presented at the 8th IAEA Technical Meeting on Fusion Power Plant Safety held in Vienna, Austria, 10–13 July 2006.

Following recommendation from the International Fusion Research Council, the IAEA organizes Technical Meetings on Fusion Safety with the aim to bring together experts to discuss the ongoing work, share new ideas and outline general guidance and recommendations on different issues related to safety and environmental (S&E) aspects of fusion research and power facilities. Previous meetings in this series were held in Vienna, Austria (1980), Ispra, Italy (1983), Culham, UK (1986), Jackson Hole, USA (1989), Toronto, Canada (1993), Naka, Japan (1996) and Cannes, France (2000). The recognized progress in fusion research and technology over the last quarter of a century has boosted the awareness of the potential of fusion to be a practically inexhaustible and clean source of energy.

The decision to construct the International Thermonuclear Experimental Reactor (ITER) represents a landmark in the path to fusion power engineering. Ongoing activities to license ITER in France look for an adequate balance between technological and scientific deliverables and complying with safety requirements. Actually, this is the first instance of licensing a representative fusion machine, and it will very likely shape the way in which a more common basis for establishing safety standards and policies for licensing future fusion power plants will be developed.

Now that ITER licensing activities are underway, it is becoming clear that the international fusion community should strengthen its efforts in the area of designing the next generations of fusion power plants—demonstrational and commercial. Therefore, the 8th IAEA Technical Meeting on Fusion Safety focused on the safety aspects of power facilities. Some ITER-related safety issues were reported and discussed owing to their potential importance for the fusion power plant research programmes.

The objective of this Technical Meeting was to examine in an integrated way all the safety aspects anticipated to be relevant to the first fusion power plant prototype expected to become operational by the middle of the century, leading to the first generation of economically viable fusion power plants with attractive S&E features.

After screening by guest editors and consideration by referees, 13 (out of 28) papers were accepted for publication. They are devoted to the following safety topics:

• power plant safety;

• fusion specific operational safety approaches;

• test blanket modules;

• accident analysis;

• tritium safety and inventories;

• decommissioning and waste.

The paper `Main safety issues at the transition from ITER to fusion power plants' by W. Gulden et al (EU) highlights the differences between ITER and future fusion power plants with magnetic confinement (off-site dose acceptance criteria, consequences of accidents inside and outside the design basis, occupational radiation exposure, and waste management, including recycling and/or final disposal in repositories) on the basis of the most recent European fusion power plant conceptual study.

Ongoing S&E studies within the US inertial fusion energy (IFE) community are focusing on two design concepts. These are the high average power laser (HAPL) programme for development of a dry-wall, laser-driven IFE power plant, and the Z-pinch IFE programme for the production of an economically-attractive power plant using high-yield Z-pinch-driven targets. The main safety issues related to these programmes are reviewed in the paper `Status of IFE safety and environmental activities in the US' by S. Reyes et al (USA). The authors propose future directions of research in the IFE S&E area.

In the paper `Recent accomplishments and future directions in the US Fusion Safety & Environmental Program' D. Petti et al (USA) state that the US fusion programme has long recognized that the S&E potential of fusion can be attained by prudent materials selection, judicious design choices, and integration of safety requirements into the design of the facility. To achieve this goal, S&E research is focused on understanding the behaviour of the largest sources of radioactive and hazardous materials in a fusion facility, understanding how energy sources in a fusion facility could mobilize those materials, developing integrated state-of-the-art S&E computer codes and risk tools for safety assessment, and evaluating and improving fusion facility design in terms of accident safety, worker safety, and waste disposal.

There are three papers considering safety issues of the test blanket modules (TBM) producing tritium to be installed in ITER. These modules represent different concepts of demonstration fusion power facilities (DEMO). L. Boccaccini et al (Germany) analyses the possibility of jeopardizing the ITER safety under specific accidents in the European helium-cooled pebble-bed TBM, e.g. pressurization of the vacuum vessel (VV), hydrogen production from the Be-steam reaction, the possible interconnection between the port cell and VV causing air ingress. Safety analysis is also presented for Chinese TBM with a helium-cooled solid breeder to be tested in ITER by Z. Chen et al (China). Radiological inventories, afterheat, waste disposal ratings, electromagnetic characteristics, LOCA and tritium safety management are considered. An overview of a preliminary safety analysis performed for a US proposed TBM is presented by B. Merrill et al (USA). This DEMO relevant dual coolant liquid lead–lithium TBM has been explored both in the USA and EU.

T. Pinna et al (Italy) summarize the six-year development of a failure rate database for fusion specific components on the basis of data coming from operating experience gained in various fusion laboratories. The activity began in 2001 with the study of the Joint European Torus vacuum and active gas handling systems. Two years later the neutral beam injectors and the power supply systems were considered. This year the ion cyclotron resonant heating system is under evaluation.

I. Cristescu et al (Germany) present the paper `Tritium inventories and tritium safety design principles for the fuel cycle of ITER'. She and her colleagues developed the dynamic mathematical model (TRIMO) for tritium inventory evaluation within each system of the ITER fuel cycle in various operational scenarios. TRIMO is used as a tool for trade-off studies within the fuel cycle systems with the final goal of global tritium inventory minimization.

M. Matsuyama et al (Japan) describes a new technique for in situ quantitative measurements of high-level tritium inventory and its distribution in the VV and tritium systems of ITER and future fusion reactors. This technique is based on utilization of x-rays induced by beta-rays emitting from tritium species. It was applied to three physical states of high-level tritium: to gaseous, aqueous and solid tritium retained on/in various materials.

Finally, there are four papers devoted to safety issues in fusion reactor decommissioning and waste management. A paper by R. Pampin et al (UK) provides the revised radioactive waste analysis of two models in the PPCS. Another paper by M. Zucchetti (Italy), S.A. Bartenev (Russia) et al describes a radiochemical extraction technology for purification of V-Cr-Ti alloy components from activation products to the dose rate of 10 µSv/h allowing their clearance or hands-on recycling which has been developed and tested in laboratory stationary conditions. L. El-Guebaly (USA) and her colleagues submitted two papers. In the first paper she optimistically considers the possibility of replacing the disposal of fusion power reactor waste with recycling and clearance. Her second paper considers the implications of new clearance guidelines for nuclear applications, particularly for slightly irradiated fusion materials.

In parallel to the ITER design process and in close cooperation with the designers a fusion specific safety approach was developed and implemented. Detailed safety assessments have been performed and documented in the ITER Generic Site Safety Report (GSSR). Following the decision on ITER construction in France, results from the GSSR and from ongoing safety related activities tailored to the Cadarache site and the French licensing process are now being used to write the ITER Preliminary Safety Analysis Report.

In the most recent European fusion power plant conceptual study the inherent fusion favourable features have been exploited, by appropriate design and choice of materials, to provide major safety and environmental advantages. The study focused on five power plant models, which are illustrative of a wider spectrum of possibilities. These span a range from relatively near-term concepts, based on limited technology and plasma physics extrapolations, to a more advanced conception. All five PPCS plant models differ substantially in their plasma physics, blanket and divertor technology, size, fusion power and materials compositions, and these differences lead to differences in the economic performance and in the details of safety and environmental impacts.

The paper uses the quite detailed information available from ITER safety documents and highlights the differences between ITER and future fusion power plants. The main areas investigated are releases and doses during normal operation and under accidental conditions, occupational radiation exposure and optimization and waste management, including recycling and/or final disposal in repositories.

This paper presents an overview of recent progress in the area of inertial fusion energy (IFE) safety and environment (S&E) in the US. Over the past several years, a significant effort has been devoted towards the development of S&E analyses for future IFE power plants. We have completed the safety assessment of various baseline IFE power plant concepts, including simulation of accident scenarios and accident consequences analyses, S&E studies of candidate target materials and discussions on waste management issues. The results from this work have allowed for a better understanding of the behaviour of radioactive sources and hazardous materials within the IFE power plant, identification of the energy sources that could mobilize those materials in case of an accident and assessment of waste management options for IFE. Currently, ongoing S&E studies for IFE are focusing on emerging design concepts, which include support to the high average power laser (HAPL) program for development of a dry-wall, laser-driven IFE power plant and collaboration with the Z-pinch IFE program for the production of an economically attractive power plant using high-yield Z-pinch-driven targets. In this paper, the main safety issues related to the HAPL and Z-IFE programs are reviewed, some recent safety highlights are presented and future directions in the IFE S&E area are proposed.

The US Fusion Program has long recognized that the safety and environmental (S&E) potential of fusion can be attained by prudent materials selection, judicious design choices and integration of safety requirements into the design of the facility. To achieve this goal, S&E research is focused on understanding the behaviour of the largest sources of radioactive and hazardous materials in a fusion facility, understanding how energy sources in a fusion facility could mobilize those materials, developing integrated state-of-the-art S&E computer codes and risk tools for safety assessment and evaluating S&E issues associated with current fusion designs. In this paper, recent accomplishments are reviewed and future directions outlined.

This paper presents a review of safety studies for accidental sequences in the European solid breeder test blanket module (TBM) system. These studies are the starting point for the Preliminary Safety Analysis Report of ITER, under preparation to get the construction permit first and then later the operation licence. In general the reduced inventory of activation products and tritium associated with the TBM system makes the impact of this test system almost negligible on the overall safety risk of ITER. Nevertheless, the possibility of jeopardizing the ITER safety concept has been analysed in connection to the consequences of specific accident sequences, e.g. the pressurization of the vacuum vessel due to the He coolant blow-down, the hydrogen production from the Be-steam reaction, the possible interconnection between the port cell and the vacuum vessel causing air ingress and the necessity to assure heat removal in the short and long periods. In the frame of this assessment, three LOCA sequences have been selected as representative of accidents judged to cover all scenarios envisaged in Cat II to IV events involving the TBM, namely, in-vessel LOCA, ex-vessel LOCA and in-box LOCA.

Safety analyses have been performed for the Chinese ITER test blanket module (TBM) design with helium-cooled solid breeder (HCSB) concept for testing in a ITER device based on the design developed with the China blanket programme. Some safety issues associated with this TBM design are explored. In particular, radiological inventories, afterheat, waste disposal ratings (WDRs), electromagnetic analysis, LOCA accident analysis and tritium safety management for the CH HCSB TBM design are given. The analysis results show that the total radiological inventory in the TBM at shutdown is very low and drops rapidly within a minute due to the decay of the Mn-56 isotope, the total afterheat is as low as 8.77 × 10⁻³ MW, which is attributed mainly to the structure and the WDRs for Ni-59, Zr-93 and Nb-94 are 1.48 × 10⁻², 4.72 × 10⁻⁹ and 1.034 × 10⁻³, respectively, and according to the US 10CFR61 regulation, these materials are qualified for shallow land burial. Results show that this test blanket concept is feasible within the existing domestic technologies and has good safety characteristics. In addition, future directions in the safety and environmental area are proposed.

The US is proposing a prototype of a dual coolant liquid lead–lithium DEMO blanket concept for testing in the International Thermonuclear Experimental Reactor (ITER) as an ITER test blanket module (TBM). Because safety considerations are an integral part of the design process to ensure that this TBM does not adversely impact the safety of ITER, a safety assessment has been conducted for this TBM and its ancillary systems as requested by the ITER project. Four events were selected by the ITER international team (IT) to address specific reactor safety concerns, such as vaccum vessel (VV) pressurization, confinement building pressure build-up, TBM decay heat removal capability, tritium and activation products release from the TBM system and hydrogen and heat production from chemical reactions. This paper summarizes the results of this safety assessment conducted with the MELCOR computer code.

The objective of this ongoing activity is to develop a fusion specific component failure database with data coming from operating experiences gained in various fusion laboratories. The activity began in 2001 with the study of the Joint European Torus (JET) Vacuum and Active Gas Handling Systems. Two years later the neutral beam injectors (NBI) and the power supply (PS) systems were considered and since last year the ion cyclotron resonant heating system is under evaluation.

The number of failures/malfunctions that have occurred during the years of operations, failure modes and, where possible, causes and consequences of the failures were identified, as well as the sets of components under analysis. Components were classified and counted in order to determine component counts by type of component, related total operating hours and related demands to operate (for components operating in an intermittent manner). Main reliability parameters (such as the failure rate and corresponding standard errors and confidence intervals) for the component types were also estimated. In this paper the NBI and PS results are presented.

Within the tritium plant of ITER a total inventory of about 2–3 kg will be necessary to operate the machine in the DT phase. During plasma operation, tritium will be distributed in the different sub-systems of the fuel cycle. A tool for tritium inventory evaluation within each sub-system of the fuel cycle is important with respect to both the process of licensing ITER and also for operation. It is very likely that measurements of total tritium inventories may not be possible for all sub-systems; however, tritium accounting may be achieved by modelling its hold-up within each sub-system and by validating these models in real-time against the monitored flows and tritium streams between the sub-systems. To get reliable results, an accurate dynamic modelling of the tritium content in each sub-system is necessary. A dynamic model (TRIMO) for tritium inventory calculation reflecting the design of each fuel cycle sub-systems was developed.

The amount of tritium needed for ITER operation has a direct impact on the tritium inventories within the fuel cycle sub-systems. As ITER will function in pulses, the main characteristics that influence the rapid tritium recovery from the fuel cycle as necessary for refuelling are discussed.

The confinement of tritium within the respective sub-systems of the fuel cycle is one of the most important safety objectives. The design of the deuterium/tritium fuel cycle of ITER includes a multiple barrier concept for the confinement of tritium. The buildings are equipped with a vent detritiation system and re-circulation type room atmosphere detritiation systems, required for tritium confinement barrier during possible tritium spillage events. Complementarily to the atmosphere detritiation systems, in ITER a water detritiation system for tritium recovery from various sources will also be operated.

To establish a new technique for monitoring and control of high-level tritium in a fuel circulation system of the fusion reactors, the applicability of a new technique to elemental tritium and tritiated water was discussed. The new technique is based on the utilization of x-rays induced by β-rays from tritium, and it is called β-ray-induced x-ray spectrometry. The applicability of the same technique to the tritium species retained in solid materials was also discussed. It was concluded from applications to gaseous and aqueous tritium under the static state that the new technique is applicable as an in-line monitor in the fuel circulation system. In addition, it was also shown that the new technique plays an important role in non-destructive evaluation of tritium retained in solid materials such as plasma-facing materials.

A sound approach to the recycling of fusion irradiated material is being developed. Study of industry experience, and consideration of realistic processing routes and techniques, provide a more sensible estimation of recycling feasibility than earlier studies based on purely radiological criteria. Under this approach, the analysis of active material in two models of the power plant conceptual study (PPCS) has been revised in more detail and accounting for the latest design features, nuclear data and international guidelines. A careful inventory of the materials has been performed, and estimation made of the radiological characteristics of all PPCS tokamak components, for the first time studying individual constituents and materials. Evaluation has been made of time scales for the radioactivity to decay to predetermined levels, which represent the spectrum of technological difficulties posed by the nature of the irradiated material. Three main mechanisms for the optimization of the materials management strategy have been identified during the assessments: segregation of components into individual materials, in situ refurbishment and stringent impurity control.

The reduction of long-term radioactivity is analysed here in V–Cr–Ti alloys, one of the proposed structural materials for fusion power plants. In particular is explored the possibility of recycling within the nuclear industry and of clearance, that is declassification to non-active material. Vanadium alloys have the potential to reach the dose rate hands-on recycling limit when used in a blanket, if noxious radioactive products coming from impurities activation are eliminated. Clearance is also possible in principle, but only if further separation of activation products of titanium is carried out after service.

The development of commercial fusion plants includes the demonstration that the waste burden for future generations would be avoided. In order to minimize the radioactive material requiring long-term storage, maximum emphasis should be placed on recycling and clearance, avoiding geologic disposal. Clearance is the most desirable option for components like the vessel, shield, magnets and bioshield, which make up a great part of the mass and are only mildly irradiated. Recently, the International Atomic Energy Agency, the US Nuclear Regulatory Commission and other institutions have revised clearance guidelines for nuclear applications. In this paper, the implications of these new standards, particularly for slightly irradiated fusion materials, are considered. The amount of clearable materials is lower and/or the cooling period is longer than previously estimated. It becomes more evident now that improvements to the clearance data are needed to include many missing fusion-specific nuclides, and that the control and measurement of impurity levels, even for materials subjected to low-neutron exposure, is important. Segregation of components into more basic parts also plays a key role in minimizing the amount of waste. Finally, the issue of public acceptability of cleared materials is discussed since the unrestricted release of radioactive materials to the market is still controversial, despite the extremely low level of activity (⩽10 µSv/y)—way below the natural background.

The wealth of experience accumulated over the past 30–40 years of fusion power plant studies must be forged into a new strategy to reshape all aspects of handling the continual stream of radioactive materials during operation and after power plant decommissioning. With tighter environmental controls and the political difficulty of building new repositories worldwide, the disposal option could be replaced with more environmentally attractive scenarios, such as recycling and clearance. We applied the three scenarios to the most recent ARIES compact stellarator power plant. All ARIES-CS components qualify as Class A or C low-level waste, according to the US guidelines, and can potentially be recycled using conventional and advanced remote handling equipment. Approximately 80% of the total waste can be cleared for reuse within the nuclear industry or, preferably, released to the commercial market. This paper documents the recent developments in radwaste management of nuclear facilities and highlights the benefits and challenges of disposal, recycling and clearance.

The issue of turbulent pinches induced by inhomogeneities in the magnetic field confining tokamak plasmas is studied in the framework of the self-consistent action-angle transport theory. It is found that, in magnetic turbulence, electron particle and energy fluxes proportional to the magnetic shear and originating from the toroidal frequency of the particle motion are different from zero only if the electron temperature is different from the ion temperature, independent of whether the electrons are trapped or untrapped. Restricting to passing electrons, it is shown that both the magnitude and the sign of these fluxes depend crucially on the ratio T_i/T_e.

The use of electrostatic probes as a diagnostic tool of the dust particles in the tokamak edge plasmas is investigated. Probe measurements of electrostatic fluctuations in the scrape-off layer of the Frascati Tokamak Upgrade revealed that some features of the signals can be explained only by a local non-propagating phenomenon. These signal features are shown to be both in qualitative and quantitative agreement with ionization, and consequent extra charge collected by the probes, due to the impact of micrometre-sized dust at a velocity of the order of 10 km s⁻¹. Electron microscope analysis of the probe surface yielded direct support for such an interpretation.

The loss of fast (i.e. suprathermal) ions from a magnetically confined fusion plasma due to the interaction with magnetohydrodynamic (MHD) instabilities has been experimentally characterized and interpreted by means of a numerical model. It is found that for a special class of instabilities, the so-called neoclassical tearing modes, fast ions losses are increased and modulated at the same frequency of the mode. This new experimental finding is explained as a result of the drift islands formed by energetic ions in particle phase space. An eventual overlapping of these drift islands leads to an orbit stochasticity and therefore to an enhancement of the fast ion losses. This explanation is confirmed by statistical analysis of simulations of fast ions trajectories performed with the ORBIT code. The mechanism is of general importance for understanding the interaction between MHD modes and fast particles in magnetic confinement experiments.

The NSTX operates at low aspect ratio (R/a ∼ 1.3) and high beta (up to 40%), allowing tests of global confinement and local transport properties that have been established from higher aspect ratio devices. The NSTX plasmas are heated by up to 7 MW of deuterium neutral beams with preferential electron heating as expected for ITER. Confinement scaling studies indicate a strong B_T dependence, with a current dependence that is weaker than that observed at higher aspect ratio. Dimensionless scaling experiments indicate a strong increase in confinement with decreasing collisionality and a weak degradation with beta. The increase in confinement with B_T is due to reduced transport in the electron channel, while the improvement with plasma current is due to reduced transport in the ion channel related to the decrease in the neoclassical transport level. Improved electron confinement has been observed in plasmas with strong reversed magnetic shear, showing the existence of an electron internal transport barrier (eITB). The development of the eITB may be associated with a reduction in the growth of microtearing modes in the plasma core. Perturbative studies show that while L-mode plasmas with reversed magnetic shear and an eITB exhibit slow changes in across the profile after the pellet injection, H-mode plasmas with a monotonic q-profile and no eITB show no change in this parameter after pellet injection, indicating the existence of a critical gradient that may be related to the q-profile. Both linear and non-linear simulations indicate the potential importance of electron temperature gradient (ETG) modes at the lowest B_T. Localized measurements of high-k fluctuations exhibit a sharp decrease in signal amplitude levels across the L–H transition, associated with a decrease in both ion and electron transport, and a decrease in calculated linear microinstability growth rates across a wide k-range, from the ion temperature gradient/TEM regime up to the ETG regime.

The impact of plasma shaping on electron heat transport is investigated in TCV L-mode plasmas. The study is motivated by the observation of an increase in the energy confinement time with decreasing plasma triangularity which may not be explained by a change in the temperature gradient induced by changes in the geometry of the flux surfaces. The plasma triangularity is varied over a wide range, from positive to negative values, and various plasmas conditions are explored by changing the total electron cyclotron (EC) heating power and the plasma density. The mid-radius electron heat diffusivity is shown to significantly decrease with decreasing triangularity and, for similar plasma conditions, only half of the EC power is required at a triangularity of −0.4 compared with +0.4 to obtain the same temperature profile. Besides, the observed dependence of the electron heat diffusivity on the electron temperature, electron density and effective charge can be grouped in a unique dependence on the plasma effective collisionality. In summary, the electron heat transport level exhibits a continuous decrease with decreasing triangularity and increasing collisionality. Local gyro-fluid and global gyro-kinetic simulations predict that trapped electron modes are the most unstable modes in these EC heated plasmas with an effective collisionality ranging from 0.2 to 1. The modes stability dependence on the plasma triangularity is investigated.

For the first time, tritium production rates in the water cooled pebble bed blanket are experimentally examined by using DT neutrons with two partial mockups; multi-layered mockup with water and pebble bed mockup. Tritium production rates (TPRs) are calculated by numerical analyses using Monte Carlo code MCNP-4C with nuclear data libraries FENDL-2.1 and JENDL-3.3. For experimental analysis on the pebble bed mockup, precise modelling method is proposed using the hexagonal close-packed heterogeneous geometry. Prediction uncertainties of TPRs are clarified through the experimental analyses. Ratios of the calculation results to the experimental results on local TPRs are 0.89–1.10 in the multi-layered mockup, and 0.95–1.13 in the pebble bed mockup. The calculation results on integrated tritium productions agree well with the experimental ones in both mockups. It is clarified that integrated tritium productions can be accurately evaluated.

Experiments to investigate transport properties under the influence of the dynamic ergodic divertor (DED) on TEXTOR are discussed. Relativistic runaway electrons are applied for studying transport properties of ergodization such as enhanced runaway loss. The ergodization causes an enhanced loss rate; this loss is higher for low relativistic electrons than for highly relativistic ones, in good agreement with particle orbit mapping. Edge transport can be controlled by the DED perturbation: in limiter H-mode plasmas ELM-like particle and heat bursts associated with the formation of enhanced edge pressure gradients are mitigated in the 6/2 configuration on the expense of a reduced pedestal height. Finally, the plasma is driven back to L-mode under the influence of the magnetic perturbation. In the 3/1 configuration the onset of tearing modes limits the possibility to affect edge transport. A mode of spontaneous density built-up has been found for the TEXTOR-DED as well. This mode is in particular strong for an inward shifted plasma; the built-up has a resonant character with respect to q(a). Langmuir probe measurements with two probe arrays show a strong influence of the magnetic ergodization on both the edge plasma equilibrium and fluctuation parameters. In particular, in the ergodic zone the turbulence properties and turbulence-driven flux are profoundly modified.

Pedestal and global plasma parameters are compared in conventional ELMy H-modes and improved confinement discharges from ASDEX Upgrade (AUG), DIII-D, JET and JT-60U with varying net input power. Both electron and ion pedestal pressures are studied. The pedestal top pressure p^PED increases moderately with power in all tokamaks, in broad agreement with the power dependence of the IPB98(y, 2) scaling. Higher pedestal pressures are observed in AUG improved H-modes and in JT-60U high β_pol discharges at q₉₅ ∼ 6.5 and high triangularity. For all machines and all scenarios a robust correlation between the total and the pedestal thermal stored energy is observed, with the ratio of the two varying between ∼0.3 and 0.5. However the relative importance of pedestal and core confinement varies from regime to regime. In AUG the confinement improvement with respect to the IPB98(y, 2) scaling is due to improved pedestal confinement in improved H-modes with early heating and to both improved pedestal and core confinement in improved H-modes with late heating. In DIII-D hybrid discharges the increase in confinement factor compared with conventional H-modes is due to improved confinement in the plasma core. JT-60U reversed shear H-modes have strong internal transport barriers and thus improved core performance. In all four tokamaks improved edge stability is correlated with increasing total β_pol and H98(y,2) increases with pedestal β_pol. The analysed multimachine data set supports a scaling expression for the pedestal stored energy derived under the assumption that the dominant loss term for the pedestal is by thermal conduction in the edge transport barrier region.

Higher moments of the plasma shape than triangularity are found to significantly affect the pedestal pressure and the edge localized mode (ELM) characteristics in DIII-D. The shape dependence of the pedestal pressure was experimentally examined by varying the squareness in the proposed ITER configuration while holding the triangularity fixed. Over this scan the pedestal pressure increased by ∼50% from highest squareness to lowest squareness. The variation of pedestal energy is found to be consistent with the stability analysis of the measured profiles. The ELM energy also varied with the shape to maintain a nearly constant fraction of the pedestal energy. Stability analysis using model shapes and pressure profiles indicates that much of the advantage of high triangularity for high pedestal pressure can be achieved in lower triangularity shapes by optimizing squareness and/or the distance of the secondary upper separatrix from the primary separatrix. In high beta discharges an increase in pedestal pressure is observed with higher global stored energy. The greatest pedestal pressure increase is at low squareness due to an increase in both the pressure gradient stability limit and the width of the pedestal. The variation in pedestal pressure with squareness was also used to optimize 'hybrid' discharges in DIII-D where a lower pedestal pressure was required for an improved overall performance. In the 'hybrid' regime low squareness resulted in a high pedestal pressure with large infrequent ELMs that eventually triggered an internal 2/1 tearing mode that locked, resulting in a disruption. At higher squareness the pedestal pressure was reduced with smaller and more rapid ELMs, resulting in the maintenance of a steady beneficial internal 3/2 tearing mode and good confinement. For all the cases studied, an increase in the pedestal width at low squareness appears to be a significant factor in the increase in the total pedestal pressure.

The error field diagnostics based on magnetic measurements outside the plasma is discussed. The analysed methods rely on measuring the plasma dynamic response to the finite-amplitude external magnetic perturbations, which are the error fields and the pre-programmed probing pulses. Such pulses can be created by the coils designed for static error field correction and for stabilization of the resistive wall modes, the technique developed and applied in several tokamaks, including DIII-D and JET. Here analysis is based on the theory predictions for the resonant field amplification (RFA). To achieve the desired level of the error field correction in tokamaks, the diagnostics must be sensitive to signals of several Gauss. Therefore, part of the measurements should be performed near the plasma stability boundary, where the RFA effect is stronger. While the proximity to the marginal stability is important, the absolute values of plasma parameters are not. This means that the necessary measurements can be done in the diagnostic discharges with parameters below the nominal operating regimes, with the stability boundary intentionally lowered. The estimates for ITER are presented. The discussed diagnostics can be tested in dedicated experiments in existing tokamaks. The diagnostics can be considered as an extension of the 'active MHD spectroscopy' used recently in the DIII-D tokamak and the EXTRAP T2R reversed field pinch.

Calculations are presented for two shots in the W7AS stellarator which differ only in the magnitude of the current in the divertor control coil, but have very different values of experimentally attainable β (⟨β⟩ ≈ 2.7% versus ⟨β⟩≈ 1.8%). Equilibrium calculations find that a region of chaotic magnetic field line trajectories fills approximately the outer 1/3 of the cross-section in each of these configurations. The field lines in the stochastic region are calculated to behave as if the flux surfaces are broken only locally near the outer midplane and are preserved elsewhere. The calculated magnetic field line diffusion coefficients in the stochastic regions for the two shots are consistent with the observed differences in the attainable β, and are also consistent with the differences in the reconstructed pressure profiles.

A numerical modelling of the dynamics of an edge-localized mode (ELM) crash in the spherical tokamak is proposed by means of a three-dimensional nonlinear magnetohydrodynamic simulation. The simulation result shows a crash of the edge pressure profile due to the spontaneous growth of the ballooning mode instability. The simulation result shows good agreement in several characteristic features of the experimental observation of large scale ELMs in an appropriate time scale: (1) relation to the ballooning instability, (2) intermediate-n precursors, (3) low-n structure on the crash, (4) formation and separation of the filament and (5) considerable amount of convective loss of plasma, where n is the toroidal mode number. Furthermore, the model is verified by examining the effect of diamagnetic stabilization and by comparing the nonlinear behaviour with that of the peeling modes. The ion diamagnetic drift terms are found to stabilize some specific components linearly; nevertheless they are not so effective in the nonlinear dynamics such as the filament formation and the amount of loss. For the peeling mode case, no prominent filament structure is found to appear in contrast to the ballooning mode case.

Radially localized electron cyclotron heating (ECH) at the 2nd harmonic (X2) is used to tailor the current profile through the modification of the temperature, and hence conductivity, profiles. The optimal conditions for current profile broadening are investigated and applied to decrease the plasma internal inductance at constant plasma current. Highly elongated plasmas are thus created at low current, where they are vertically unstable with peaked Ohmic current profiles, using active current profile broadening by far off-axis X2 ECH to ensure the vertical stability of the plasma. Central 3rd harmonic (X3) ECH is also applied and its impact on the current profile is described. The magnetohydrodynamic activity and in particular the replacement of the sawtooth by a multi-harmonics mode, resonant on the q = 1 surface and observed at high density and low internal inductance, is investigated.

Recently gamma(γ)-ray emission measurements have been performed for the first time in trace tritium experiments in the Joint European Torus (JET) [Kiptily et al 2004 Phys. Rev. Lett.93 115001]. The decay in the γ-ray emission has been measured after a short duration tritium neutral beam injection (NBI) into deuterium plasmas. The measured γ-ray emission is produced in the reaction between the α-particles and intrinsic beryllium impurity, and therefore it carries information about the α-particle behaviour. The interpretation of the γ-ray emission decay after the tritium NBI and short single gas puff into H-mode plasmas, based on the modelling of the α-particle evolution with the TRANSP code, is given here. It is shown that the α-particle confinement may determine the decay time of the γ-ray emission during the very short post-NBI phase (∼150–200 ms) unresolved with present diagnostics (time resolution of γ-ray emission measurements is 250 ms). The later γ-ray decay correlates with the confinement of thermalized tritium (i.e. with the decay of the α-particle source). An upper estimate of the α-particle confinement time is determined, but this estimate exceeds the α-particle slowing down time giving a large uncertainty in the estimation of anomalous (non-collisional) losses in these plasmas. The α-particle behaviour is analysed also in discharges with tritium gas puffs with the goal of finding a scenario where the α-particle production and losses could be clearly distinguished. It is found that a large time delay between the decay of the α-particle source and its products occurs in discharges performed at high plasma density due to the large transient orbit losses of α-particles. These scenarios are proposed for testing α-particle orbit losses in future deuterium–tritium experiments.

Theoretical and computational investigations of boundary-plasma microturbulence which take into account important effects of the geometry of diverted tokamaks—in particular, the effect of X-point magnetic shear and the termination of field lines on divertor plates—are presented. We first generalize our previous 'heuristic boundary condition' which describes, in a lumped model, the closure of currents in the vicinity of the X-point region to encompass three current-closure mechanisms. We then use this boundary condition to derive the dispersion relation for low-beta flute-like modes in the divertor-leg region under the combined drives of curvature, sheath impedance and divertor tilt effects. The results indicate the possibility of strongly growing instabilities, driven by sheath boundary conditions, and localized in either the private or common flux region of the divertor leg depending on the radial tilt of divertor plates. We revisit the issue of X-point effects on blobs, examining the transition from blobs terminated by X-point shear to blobs that extend over both the main SOL and divertor legs. We find that, for a main-SOL blob, this transition occurs without a free-acceleration period as previously thought, with X-point termination conditions applying until the blob has expanded to reach the divertor plate. We also derive propagation speeds for divertor-leg blobs. Finally, we present fluid simulations of the C-Mod tokamak from the BOUT edge fluid turbulence code, which show main-SOL blob structures with similar spatial characteristics to those observed in the experiment, and also simulations which illustrate the possibility of fluctuations confined to divertor legs.

Plasma momentum transport within magnetic surfaces plays a fundamental role in a number of toroidal plasma physics issues, such as turbulence suppression, impurity transport, bootstrap current generation and the shielding of resonant magnetic error field perturbations. Stellarators provide opportunities for improved understanding of plasma flow effects because (a) new forms of quasi-symmetry (e.g. helical, poloidal) can be produced that differ significantly from the tokamak and (b) symmetry-breaking effects (always present to some degree) reduce the close coupling between parallel and cross-field transport characteristics of symmetric systems. External control coils can also be used to further enhance or suppress such effects. A method has been developed to evaluate the variation of neoclassical self-generated plasma flows in stellarators both within and across magnetic surfaces. This introduces a new dimension into both the optimization of stellarators and to the improved understanding of the existing confinement database. Application of this model to a range of configurations indicates that flow directionality and shearing rates are significantly influenced by the magnetic structure. In addition, it is demonstrated that flows in stellarators are sensitive to profile effects and the presence of external momentum sources, such as neutral beams.

The lowest aspect ratio possible in a spherical tokamak is defined by limiting the plasma on its centre column, which might therefore maximize many physics benefits of this fusion approach. A key issue for such discharges is whether loads exhausted onto the small surface area of the column remain acceptable. A first series of centre-column-limited pulses has been examined on MAST using fast infra-red thermography to infer incident power densities as neutral-beam heating was scanned from 0 to 2.5 MW. Simple mapping shows that efflux distributions on the column armour are governed mostly by magnetic geometry, which moreover spreads them advantageously over almost the whole vertical length. Hence steady peak power densities between sawteeth remained low, <1 MW m⁻², comparable with the target strike-point value in a reference diverted plasma at lower power. Plasma purity and normalized thermal energy confinement through the centre-column-limited (CCL) series were also similar to properties of MAST diverted cases. A major bonus of CCL geometry is a propensity for exhaust to penetrate through its inner scrape-off layer connecting to the column into an expanding outer plume, which forms a 'natural divertor'. Effectiveness of this process may even increase with plasma heating, owing to rising Shafranov shift and/or toroidal rotation. A larger CCL device could potentially offer a simpler, more economic next-step design.

Active control of the resistive wall mode has been considered as an alternative way, besides passive control by plasma rotation, to stabilize the mode in advanced scenarios in ITER. We show that a significant improvement of the feedback system is achievable by optimizing the detection system, without changing the active coil system in the design. Both analytical theory in cylindrical geometry, and toroidal calculations for ITER equilibria, show that the resistive wall mode can be stabilized for plasma pressures close to the ideal-wall beta limit, using the new technique with even radial sensors. This high plasma pressure is stabilized with the feedback coils outside the vacuum vessel, as in the nominal ITER design.

A non-linear MHD code, named JOREK, is under development with the aim of studying the non-linear evolution of the MHD instabilities thought to be responsible for edge localized modes (ELMs): external kink (peeling) and medium-n ballooning modes. The full toroidal X-point geometry is taken into account including the separatrix, open and closed field lines. Analysis of the influence of the separatrix shows a strong stabilization of the ideal and resistive MHD external kink/peeling modes. One instability remains unstable in the presence of the X-point, characterized by a combination of a tearing and a peeling mode. The so-called peeling–tearing mode shows a much weaker dependence on the edge q. Non-linearly the n = 1 peeling–tearing mode saturates at a constant amplitude yielding a mostly kink-like perturbation of the boundary with an island-like structure close to the X-point. The non-linear evolution of a medium-n ballooning mode shows the formation of density filaments. The density filaments are sheared off from the main plasma by an n = 0 flow non-linearly induced by the Maxwell stress. The amplitude of the ballooning mode is limited by this n = 0 flow and multiple (in time) density filaments can develop to bring the plasma below the stability boundary.

Fluctuations and particle transport in the scrape-off layer of TCV plasmas have been investigated by probe measurements and direct comparison with two-dimensional interchange turbulence simulations at the outer midplane. The experiments demonstrate that with increasing line-averaged core plasma density, the radial particle density profile scale length becomes broader. The particle and radial flux density statistics in the far scrape-off layer exhibit a high degree of statistical similarity with respect to changes in the line-averaged density. The plasma flux onto the main chamber wall at the outer midplane scales linearly with the local particle density, suggesting that the particle flux here can be parameterized in terms of an effective convection velocity. Experimental probe measurements also provide evidence for significant parallel flows in the scrape-off layer caused by ballooning in the transport of particles and heat into the scrape-off layer. The magnitude of this flow estimated from pressure fluctuation statistics is found to compare favourably with the measured flow offset derived by averaging data obtained from flow profiles observed in matched forward and reversed field discharges. An interchange turbulence simulation has been performed for a single, relatively high density case, where comparison between code and experiment has been possible. Good agreement is found for almost all aspects of the experimental measurements, indicating that plasma fluctuations and transport in TCV scrape-off layer plasmas are dominated by radial motion of filamentary structures.

A low frequency torsional Alfven wave (angular frequency ω ∼ 0.3ω_ci0, ω_ci0; ion gyro frequency in vacuum) is induced in an extremely high beta (beta = plasma pressure/pressure of confining magnetic field) plasma with the field reversed configuration (FRC) by an external antenna. When the wave is introduced from outside towards the magnetic axis of the FRC plasma, not only the Alfven velocity v_A but also the local ion gyro frequency ω_ci decreases as the density increases and the magnetic field decreases. The spatial structure of the wave is examined in the region between the separatrix and the location near the O-point. Near the former location, the plasma is tenuous and the magnetic field has nearly external value so that ω ∼ 0.3ω_ci0, and near the latter location, the beta value is large and ω > 0.7ω_ci. The observed amplitude of the toroidal component of the magnetic disturbance has a broad peak around the radial location of the resonance position obtained from the cold plasma theory, although the cold plasma theory yields a sharper spatial structure. The polarization of the magnetic disturbance component perpendicular to the static magnetic field is left-handed outside the theoretical resonance position and right-handed deep in the plasma. The measured perpendicular wave number k_r is small except for the region near the resonance position where the kinetic Alfven wave is expected to have a large wave number. This behaviour will be interpreted as, with the help of the warm plasma theory, the torsional Alfven wave evolves into the kinetic Alfven wave and is then converted to the compressional wave with a smaller wave number when the wave propagates inside the plasma.

An integrated simulation code TOPICS-IB based on a transport code with a stability code for the peeling–ballooning modes and a scrape-off-layer (SOL) model has been developed to clarify self-consistent effects of edge localized modes (ELMs) and the SOL on the plasma performance. Experimentally observed collisionality dependence of the ELM energy loss is found to be caused by both the edge bootstrap current and the SOL transport. The bootstrap current decreases with an increase in collisionality and intensifies the magnetic shear at the pedestal region. The increase in the magnetic shear reduces the width of eigenfunctions of unstable modes, which results in the reduction of both the area of the ELM enhanced transport and the ELM enhanced transport near the separatrix. On the other hand, when an ELM crash occurs, the energy flows into the SOL and the SOL temperature rapidly increases. The increase in the SOL temperature lowers the ELM energy loss due to the flattening of the radial edge gradient. The parallel electron heat conduction determines how the SOL temperature increases. For higher collisionality, the conduction becomes lower and the SOL electron temperature increases more. By the above two mechanisms, the ELM energy loss decreases with increasing collisionality.

After installation of ferritic steel tiles, fast ion losses due to toroidal field ripple have been reduced by 1/2–1/3. The increase in absorbed power at same injection power can reduce the required number of neutral beam injector (NBI) units to sustain a given normalized beta, β_N, resulting in a better flexibility of torque input by increasing the available combination of tangential NBI units. By making use of these advantages to sustain an internal transport barrier (ITB), the performance of long-pulse ELMy H-mode plasmas was improved in terms of sustained duration time for both high β_N and high thermal confinement enhancement factor (H_H98(y,2)). High β_N > 2.3 together with H_H98(y,2) ∼ 1 was sustained for 23.1 s (∼12τ_R, where τ_R is the current diffusion time) at q₉₅ ∼ 3.3, which also provide high β_NH_H98(y,2) ≥ 2.2 and a bootstrap current fraction of ≥40%. β_NH_H98(y,2) of 2.0 was sustained for 28.6 s, which is limited by the maximum injection period of 30s for NBI system. These long-pulse plasmas are possible candidates for ITER hybrid operation scenario. Improved confinement is characterized by the larger thermal components at a given density maintained by lower heating power than in previous experiments. The strength of the ITB depends on the pedestal temperature, which varies with edge density while keeping constant the edge pressure (limited by type I ELMs). The fact that co-toroidal rotation as a result of reduced fast ion losses provides better quality of T_e-ITB also contributes the improvement of thermal plasma confinement. These long-pulse plasmas indicate that further investigation to establish high performance plasmas longer than the time scale of wall saturation (τ_W) with active particle control is essential to establish the operational scenarios for next step devices, where the wall pumping does not work.

The latest results of B2-Eirene modelling of ITER divertor operation are presented. The operational window is further explored with an improved model of the neutral transport, the effect of gas leaks between the divertor cassettes is assessed and the sensitivity of the results to the features of the divertor geometry and to the gas puffing arrangement is analysed. The presence of neutral–neutral collisions in the model makes the results rather insensitive to the detail of the divertor shape. The analysis shows that the effect of the gas leak on the divertor performance in ITER is weak and therefore inter-cassette sealing may be not critical.

The effects of toroidal rotation and toroidal field (TF) ripple on the edge pedestal structure and edge-localized-modes (ELMs) were examined in JT-60U. The amplitude of the TF ripple was changed by the installation of ferritic steel tiles (FSTs). The profile of toroidal rotation V_T became less counter with FSTs, particularly in the case where co-NBI was used in a large volume configuration. In this case, the plasma pressure was raised across the whole profile of the plasma. At the plasma edge, higher pedestal temperature was obtained with the growth of pedestal width. However, the effect of FSTs became less significant in a small volume configuration. As the V_T became less counter at the pedestal, ELM frequency f_ELM was reduced and ELM energy loss ΔW_ELM was increased at fixed power crossing the separatrix P_sep. The observed larger ELM energy drop at V_T in a relatively co-direction involves the ELM perturbations of the electron temperature profile across an ELM that extends radially more inward. In addition, the inter-ELM transport loss is reduced and the pedestal pressure p^ped is weakly raised. The effect of FSTs appeared clearly in the large volume configuration where p^ped is raised even at a given V_T at the pedestal while p^ped is not changed by FSTs at the small volume configuration. This suggests that the reduction of the TF ripple, the magnitude of which can affect the V_T profile, plays a role in the increasing of p^ped. For increased p^ped due to the use of co-NBI and the reduction of TF ripple, the spatial width of the pedestal ion temperature became greater.

Electron internal transport barriers (eITBs) are generated in the TCV tokamak with strong electron cyclotron resonance heating in a variety of conditions, ranging from steady-state fully noninductive scenarios to stationary discharges with a finite inductive component and finally to transient current ramps without current drive. The confinement improvement over L-mode ranges from 3 to 6; the bootstrap current fraction is invariably large and is above 70% in the highest confinement cases, with good current profile alignment permitting the attainment of steady state. Barriers are observed both in the electron temperature and density profiles, with a strong correlation both in location and in steepness. The dominant role of the current profile in the formation and properties of eITBs has been conclusively proven in a TCV experiment exploiting the large current drive efficiency of the Ohmic transformer: small current perturbations accompanied by negligible energy transfer dramatically alter the confinement. The crucial element in the formation of the barrier is the appearance of a central region of negative magnetic shear, with the barrier strength improving with increasingly steep shear. This connection has also been corroborated by transport modelling assisted by gyrofluid simulations. Rational safety-factor (q) values do not appear to play a role in the barrier formation, at least in the q range 1.3–2.3, as evidenced by the smooth dependence of the confinement enhancement on the loop voltage over a broad eITB database. MHD mode activity is however influenced by rational q values and results in a complex, sometimes cyclic, dynamic evolution.