A multichannel far-infrared laser-based POlarimeter-INTerferometer (POINT) system utilizing the three-wave technique is under development for current density and electron density profile measurements in the EAST tokamak. Novel molybdenum retro-reflectors are mounted in the inside wall for the double-pass optical arrangement. A Digital Phase Detector with 250 kHz bandwidth, which will provide real-time Faraday rotation angle and density phase shift output, have been developed for use on the POINT system. Initial calibration indicates the electron line-integrated density resolution is less than 5 × 10(16) m(-2) (∼2°), and the Faraday rotation angle rms phase noise is <0.1°.

Deuterium high-confinement (H-mode) plasmas, lasting up to 3.45 s, have been generated in the EAST by ion cyclotron range of frequency (ICRF) heating. H-mode access was achieved by coating the molybdenum-tiled first wall with lithium to reduce the hydrogen recycling from the wall. H-mode plasmas with plasma currents between 0.4 and 0.6 MA and axial toroidal magnetic fields between 1.85 and 1.95 T were generated by 27 MHz ICRF heating of deuterium plasma with hydrogen minority. The ICRF input power required to access the H-mode was 1.6–1.8 MW. The line-averaged density was in the range (1.83–2.3) × 1019 m−3. 200–500 Hz type-III edge localized mode activity was observed during the H-mode phase. The H-mode confinement factor, H98IPB(y, 2), was ∼0.7.

Experimental and modeling investigations on the Experimental Advanced Superconducting Tokamak (EAST) show attractive confinement and stability properties in fully non-inductive, high poloidal beta plasmas. In the 2018 EAST experimental campaign, extended operation regimes of steady-state scenario were obtained (β P ~ 1.9 & β N ~ 1.5 & H 98y 2 ~ 1.3 of using only RF heating) with a high bootstrap current fraction (f BS ~ 47%) and n e /n GW ~ 70%. The confinement quality, H 98y 2 ~ 1.3, is much better than standard H-mode, and stationary peaked electron temperature profiles and peaked current density profile when ~1 MW of ECH and ~2.6 MW of LHW are both deposited in the core region. The observed improvement in plasma confinement is much better (H 98y 2 ~ 1.3) when compared with the RF-dominant heating experiments in the EAST 2016-2017 experimental campaign (H 98y 2 ~ 1.1). Integrated modeling prediction suggests that high electron density would increase the plasma performance and bootstrap current fraction, which is consistent with the general experimental trend. Linear analysis shows that the high-k (k y > 1) modes instability (ETG) is suppressed in the core region. Also, the Shafranov shift is shown to play a role in the suppression of the electron turbulent energy transport. Besides the modeling predictions, the validation of the predicted of the effect of ECH on the plasma confinement in recent experiments was done and the experimental results were consistent with the modeling results. The validation results also suggest that when ECH is deposited in the core region in the RF heating experiments, increasing the ECH heating power from 0.5 MW to 1.0 MW does make a small improvement in the bootstrap current fraction. The high bootstrap fraction scenario realized on EAST and the investigation to achieve higher-performance plasma would help expanding the operation regime on EAST.

Dedicated experiments focusing on the influence of lower hybrid waves (LHWs) on edge-localized modes (ELMs) were first performed during the 2012 experimental campaign of EAST, via modulating the input power of LHWs in the high-confinement-mode (H-mode) plasma mainly sustained by ion cyclotron resonant heating. Natural ELMs are effectively mitigated (ELM frequency increases, while its intensity decreases dramatically) as the LHW is applied, observed over a fairly wide range of plasma current or edge safety factor. By scanning the modulation frequency (fm) of LHW injected power in a target plasma dominated by the so-called small ELMs, we conclude that large ELMs with markedly larger amplitudes and lower frequencies are reproduced at low modulation frequencies (fm < 100 Hz). Analysis of the evolution of edge extreme ultraviolet radiation signals further indicates that plasma fluctuations at the pedestal region indistinctively respond to rapid modulation (fm ⩾ 100 Hz) of LHW injected power. This is proposed as the mechanism responsible for the observed fm dependence of the mitigation effect induced by LHWs on large ELMs. In addition, a critical threshold of LHW input power PLHW is estimated as , beyond which the impact of applied LHWs on ELM behaviours can be achieved. Finally, Langmuir probe measurements suggest that, rather than the concentration of free energy into a narrowband quasi-coherent precursor commonly observed growing until the ELM crash, the continuous development of broadband turbulence during the ELM-absent phase with the application of LHWs might contribute to the avoidance of ELM crashes. These results present new insights into existing experiments, and also provide some foundations and references for the next-step research about exploring in more depth and improving this new attractive method to effectively control the ELM-induced very large transient heat and particle flux.

A double-pass, radially-viewing, multichannel far-infrared (FIR) polarimeter/interferometer system is under development for current density profile and electron density profile measurements in the EAST tokamak. The system utilizes three 432.5 µm CW formic acid FIR lasers pumped by three CO 2 lasers. Each of the three FIR lasers can generate high output power of more than 30 mW per cavity. Two lasers, with slight frequency offset (∼ 1 MHz), will be made collinear with counter-rotating circular polarization in order to determine the Faraday effect by measuring their phase difference. The third laser also frequency offset, will be used as a reference providing local oscillator (LO) power to each mixer so that one can obtain the phase shift caused by the plasma electron density. Novel molybdenic retro-reflectors with shutter protection have been designed and will be mounted on the inner vessel wall in EAST. The retro-reflectors can withstand baking temperature up to 350 • C and discharge duration more than 1000 s. Vibrations and path length changes due to thermal expansion will be compensated using a He-Ne interferometer as the second color. VDI planar-diode Integrated Conical Horn Fundamental Mixers optimized for high sensitivity, typical 750 V/W, will be used. Initially a five-chord system will be installed in 2013 and an eleven-chord system will be implemented on the core region of EAST plasmas. MHz frequency response allows system to resolve fast MHD events such as tearing/neoclassical tearing, disruptions and fast-particle modes. Preliminary design will be presented.

Neoclassical tearing modes (NTM) are observed in discharges with auxiliary heating LH+ICRF and LH only during H-mode in EAST. The m/n = 2/1 NTM is triggered by strongly coupling with an m/n = 1/1 internal mode. Here, LH and ICRF are the abbreviations of lower hybrid resonance heating and ion cyclotron resonance frequency heating, respectively. The mode number of the NTM is m/n = 2/1, where m is the poloidal mode number and n is the toroidal mode number. Just before the triggering of NTMs, an m/n = 1/1 internal mode appears in the soft x-ray emission at plasma centre when the intensity of hard x-ray (I HX ) reaches a critical value. The mode, characterized by frequency chirping in the spectrum, may be related to suprathermal electrons produced by LH. The saturated magnetic island width w sat of the NTM is strongly correlated with poloidal β p . Normalized β N,onset and the magnetic island critical width w crit increase with electron temperature T e .

In this study, a full convolutional neural network is trained on a large database of experimental EAST data to classify disruptive discharges and distinguish them from non-disruptive discharges. The database contains 14 diagnostic parameters from the ∼104 discharges (disruptive and non-disruptive). The test set contains 417 disruptive discharges and 999 non-disruptive discharges, which are used to evaluate the performance of the model. The results reveal that the true positive (TP) rate is ∼ 0.827, while the false positive (FP) rate is ∼0.067. This indicates that 72 disruptive discharges and 67 non-disruptive discharges are misclassified in the test set. The FPs are investigated in detail and are found to emerge due to some subtle disturbances in the signals, which lead to misjudgment of the model. Therefore, hundreds of non-disruptive discharges from training set, containing time slices of small disturbances, are artificially added into the training database for retraining the model. The same test set is used to assess the performance of the improved model. The TP rate of the improved model increases up to 0.875, while its FP rate decreases to 0.061. Overall, the proposed data-driven predicted model exhibits immense potential for application in long pulse fusion devices such as ITER.

In this study, a long short-term memory (LSTM) model is trained on a large disruption warning database to predict the disruption on EAST tokomak. To compare the performance of the proposed model with the previously reported full convolutional neural network (CNN) (Guo et al 2020 Plasma Phys. Control. Fusion 63 025008), the same data set and diagnostic signals are used. Based on the test set, the area under the receiver operating characteristic curve, i.e. the AUC value of the LSTM model is obtained as 0.87, and the true positive rate (TPR) is ~87.5%, while the false positive rate (FPR) is ~15.1%. Since the LSTM model is more sensitive to radiation fluctuations than CNN, the prediction performance of LSTM model is inferior to that of CNN model (for CNN, AUC ~0.92, TPR ~87.5%, FPR ~6.1%). However, the advance warning time of LSTM model is 14 ms earlier than that of CNN. To reduce the FPR and improve the performance of the model, more fast bolometer channels are added as the input signals of the LSTM model, including the radiation from the upper and lower edges and the plasma core. Consequently, for the same test set, the AUC value increases to 0.89, and the FPR decreases to ~9.4%, but the TPR also decreases to ~83.9%. In addition, the sensitivity of the model to radiation fluctuations caused by impurity behavior decreases significantly, and the warning time becomes 8.7 ms earlier as compared to that of the original model. Overall, it is proved that deep learning algorithms exhibit immense application potential in the disruption prediction of long-pulse fusion devices.