Modelling the Ohmic L-mode ramp-down phase of JET hybrid pulses using JETTO with Bohm–gyro-Bohm transport

Abstract

The empirical Bohm-gyro-Bohm (BgB) transport model implemented in the JETTO code is used to predictively simulate the purely Ohmic (OH), L-mode current-ramp-down phase of three JET hybrid pulses, which combine two different ramp rates with two different electron densities (at the beginning of the ramp). The modelling is discussed, namely the strategy to reduce as much as possible the number of free parameters used to benchmark the model predictions against the experimental results. Hence, keeping the gas puffing rate as measured whilst controlling the line-averaged electron density via the recycling coefficient (which in the modelling is taken at the separatrix instead of the wall), one of the many possible ways to fix the total particle source, it is shown that the BgB model reproduces well the experimental data, as far as both average quantities (plasma internal inductance and volume-averaged electron temperature) and profiles (electron density and temperature) are concerned, with relative errors remaining mostly below 20%. The sensitivenesses with respect to the recycling coefficient, the ion effective charge, the energy of neutrals entering the plasma through the separatrix and the need to introduce a particle pinch are assessed; the necessity for a proper sawtooth model if experimental results are to be reproduced is also shown. The strong non-linear coupling in a OH plasma between density, temperature and current (essentially via interplay between the powerbalance equation, Joule's heating with a temperature-dependent resistivity and the dependence of BgB transport coefficients on profile gradients) is put in evidence and analyzed in light of modelling results. It is still inferred from the modelling that the real value of the recycling coefficient at the separatrix (basically, the so-called fuelling efficiency times the actual recycling coefficient at the wall) must become close to one in the final stages of the discharges, when the gas puffing is switched off and so recycling comes to be the only source of particles. If the wall recycling remains close to one (as standard for tokamaks), this may indicate that the fuelling efficiency also approaches unity, apparently consistent with the observed fact that the plasma is pushed towards the machine wall at the end of the current ramps.

Evaluation of fuelling requirements for core density and divertor heat load control in non-stationary phases of the ITER DT 15 MA baseline scenario

“…Anomalous transport is described by the Bohm/gyro-Bohm (BgB) H-mode model [27] together with a collisionality dependent inwards pinch term. Coefficients were adapted in order to match ITER predictions obtained with the gyro-Landau fluid model GLF-23 [28] as described in [5,29] in the H-mode phases of the discharge, and by the standard Bohm/gyroBohm L-mode model [30,31] (including an inwards pinch term proportional to 0.5 • D i,BgB ) when the plasma is in L-mode. Impurity reaction crosssections are evaluated by ADAS [32].…”

Influences of heating and plasma density on impurity production and transport during the ramp-down phase of JET ILW discharge

“…a disruption rate) of 3.4%. An analysis of the pure ohmic and L-mode current-ramp-down (RD) phases of three hybrid JET pulses is shown in [3]. The JET disruption rate has dramatically increased due to an ITER-like wall (ILW), equipped with a W divertor and a Be wall [4].…”

Exact conservative solutions of fluid models for the scrape-off layer as the ancestors of plasma blobs?

“…In general, albeit more apparent in the extended solutions of figures 7 and 8 than in the more localised, blob-like solutions of figures 9 and 10, the error patterns for Ln and ∇ 2 φ tend to remain more localised than for φ , which can be explained by the fact that the logarithmic density Ln and vorticity ∇ 2 φ are governed by the two very similar equations in (2) (which are numerically solved via an RK4 scheme involving only the closest neighbouring points in the grid), whereas the potential φ is calculated by inverting Poisson's equation (which is done via a matrix inversion or elimination operation that ends up involving all points in the grid). Global relative errors have also been defined according to [33,34]…”