The development of steady state operational regimes with improved confinement and stability is known as “advanced tokamak” research. In JET advanced tokamak research mainly focuses on plasmas with internal transport barriers (ITBs), generated by modifications of the current density profile in the plasma core. New realtime measurements and feedback control algo- rithms have been developed and implemented to successfully control the ITB dynamics and the current density profile in highly noninductive regimes. Global confinement parameters, ion or electron temperatures, density and safety factor profiles (q-profiles) can now be calculated in real-time. The real time control of stationary ITBs in full current drive operation represents a major milestone towards the definition and viability of steady state tokamak. The JET real time control experiments pave the way to long steady-state ITB operation in preparation of the ITER ‘advanced’ tokamak scenario.

Safety factor “q”:
Number of turns the helical magnetic field lines in a tokamak make around the major circumference per single turn round the minor circumference. This has no connection with the ordinary sense of “safety” but relates to plasma stability.

Non-inductive current:
Toroidal plasma current not driven by the transformer. It can be generated by the injection of waves in the toroidal direction in the plasma of a tokamak. It can also be selfgenerated by the pressure gradients in the plasma: this is the so-called “bootstrap” effect.

Improvement of the tokamak concept, i.e. confinement and stability, is a crucial challenge that could lead to operating the device in a high-performance continuous mode. In a steady state tokamak reactor, like in a foreseen ‘advanced scenario’ for ITER, the plasma current will be entirely sustained by noninductive means and the self-generated bootstrap current must provide a significant fraction of the plasma current. The noninductive currents, pressure profiles, confinement and safety factor profiles (q) are strongly coupled in steady state operation with high fusion performance, as illustrated in Fig. 1. Modification of the safety factor profile will affect the confinement and core pressure (e.g. the formation of core transport barriers), and variation of the core pressure will modify the selfgenerated bootstrap current that in turns influences the safety factor profile. Maintaining the required high confinement level in steady state will therefore require active control of the plasma profiles.

In JET non-inductive quasi-stationary operation has been achieved in high fusion performance discharges with a large bootstrap current fraction, with a well developed ITB affecting both ions and electrons thermal confinement. In those experiments, simultaneous feedback control of the electron temperature gradient and of the neutron yield has allowed an ITB to be maintained with a pre-requested strength in quasisteady state for 7.5 seconds (Fig. 2).

An Interfero-polarimeter diagnostics is used to determine the plasma density and the current density profile. This is achieved by measuring the phase and polarisation shifts of two infrared Laser beams: one probing beam crossing the plasma and one reference beam.

Grad-Shafranov equation: A differential equation which describes the equilibrium of an axi-symmetric system such as a tokamak plasma.

The actuators were the ion cyclotron resonance heating (ICRH) and neutral beam injection (NBI) power, respectively. One of the main conclusions of this first set of experiments is the role played by the pre-programmed Lower Hybrid Current Drive (LHCD), during the prelude and high power heating phases, in pre-forming and sustaining the safety factor profile evolution, respectively. The transport reduction, observed also in other tokamak devices, is associated with localised turbulence suppression, which is related to the precise shape of the q-profile. Moreover, a strong correlation has also been shown between the creation of ITB and the appearance of integer-q magnetic surfaces at parti-cular locations. Active feedback control of the current density profile in addition to the pressure profile, is thus a key to reproducibly trigger an ITB and to maintain its quality and strength in a steady state condition.

Previously real-time control of the q-profile had been performed through the internal inductance parameter. Its control is not sufficient to maintain an optimised q-profile in ITB discharges. Recent efforts have been made to provide a real-time identification of the q-profile and develop an algorithm, which allows its control. The algorithm uses as input the signals of magnetic and interfero-polarimeter diagnostics. The approach described in the previous paragraph to control the pressure profile is based on decoupled control loops for the core pressure and maximum temperature gradient with devoted actuators (NBI and ICRH, respectively). For the q-profile, a ‘model-based’ control scheme was followed, in which more information on the spatial structure of the system is taken into account. To validate this ‘model based’ technique direct control of the safety factor profile has been attempted using LHCD as the only actuator. The experiment was performed during an extended LHCD prelude phase. In a tokamak like ITER, this phase would precede the application of the main additional heating power to create an ITB once the desired optimised q-profile is obtained. The plasma parameters were chosen in order to be close to those needed for a purely non-inductive regime with the available LH power, and thus have a larger flexibility for obtaining the required q-profiles. The feedback control was performed on five points of the q-profile located at fixed normalised radii. The reference q-profile is reached within 12s (Fig. 3).

To reach the pre-set reference q-profile the controller minimises in the least square sense the difference between the five target q-values and the real-time measurements. The successful and very recent experiment reported here should be considered as a ‘proof of principle’. In the near future, this general ‘model-based’ approach will be implemented to control high-pressure, high-bootstrap fraction, ITB discharges where pressure and current density profiles are strongly (non-linearly) coupled. In addition, a substantial effort is being devoted to a better identification of the magnetic equilibrium by solving the Grad-Shafranov equation in real time.

The sustainment and real time control of ITBs in full current drive operation with a significant fraction of bootstrap current represents a major milestone towards the definition and viability of steady state tokamak operation. Real time measurements of the kinetic and magnetic profiles together with ‘model-based’ feedback control algorithms will be extensively used in future experimental campaigns on JET to further increase the plasma fusion performance.