Recent Publications

Soft H-L back transitions induced by RMP coils in high q95 plasmas on EAST

Linming Shao, Hui-Hui Wang, Shouxin Wang, Zichao Lin, R Chen, Shuai Gu, Wenbo Hua, Pan Li, Yichao Li, zhiguo li, et al.

Soft H-mode to L-mode (H-L) back transitions, combined with simultaneous particle and heat pumping via n=2 resonant magnetic perturbations (RMPs), have been achieved in EAST plasmas at low safety factor (q95=3.5-4.0) without reducing auxiliary heating power [Shao L M, et al., 2024 PPCF 66 035018]. For the termination control of ITER, which is also equipped with RMP coils, this process inevitably drives the plasma toward a high q95,where the established n =2 RMP is unlikely to produce a soft H-L back transition, because the edge-resonant surface at such high q95 requires a high poloidal mode number m, and the n=2 RMP field is strongly attenuated at such high m and couples only weakly to the edge. However, an n =1 RMP can restore edge coupling, enabling reliable particle and heat pump-out without radiative collapse or current-profile MHD risks on EAST. Within an operating window constrained by impurity and NBI conditions, both the n=1 and n=2 RMPs exhibit a reproducible energy confinement time threshold at the H-L transition τ^{H-L}_{e}∼ 49 ms, which can serve as a control trigger.

Real-Time Hydrogen-Deuterium Fraction Feedback Control on EAST

Li Guo, Juan Huang, Kai Wu, Q. P. Yuan, Kedong Li, Wei Gao, Chengrui Wu, Junjie Huang, Chenyu Pan, Zixin Zhang

A simulink-to-FunctionBlock transformation framework for rapid integration of plasma control algorithms

J.Q. Zhu, Zherui Cai, Q P Yuan, Zhongmin Huang, Ruirui Zhang, Junjie Huang, Heru Guo, Gen Xu, Bingjia Xiao

To support high-performance long-pulse plasma discharges, the Lingshu Plasma Control System (PCS) has been successfully developed and deployed, shifting the focus of current research toward the development and iterative refinement of advanced control algorithms. Simulink is a widely used and powerful tool for control algorithm development. However, integrating Simulink models into the FunctionBlock-based architecture of Lingshu PCS still requires substantial manual adaptation, resulting in low integration efficiency and prolonged deployment cycles. To address this issue, this paper proposes a Simulink-to-FunctionBlock transformation framework for the Lingshu PCS. The framework establishes an automated workflow from model development to system deployment, enabling the direct conversion of Simulink models into deployable FunctionBlocks. By bridging the semantic gaps between Simulink models and the FunctionBlock architecture, the proposed approach significantly reduces manual integration effort and accelerates algorithm deployment. The proposed framework is validated using a plasma control algorithm from the EAST tokamak. Experimental results demonstrate that Simulink models can be successfully transformed into FunctionBlocks and deployed on the Lingshu PCS while preserving the original algorithm behavior and satisfying real-time control requirements, thereby validating the effectiveness of the proposed framework.IntroductionMagnetic confinement tokamaks offer a promising pathway toward the realization of clean fusion energy. In recent years, next-generation fusion facilities in China, including the China Fusion Engineering Test Reactor (CFETR)[1] and the Comprehensive Research Facility for Fusion Technology (CRAFFT/BEST), have entered an accelerated construction phase. These devices are designed to operate under more demanding conditions, featuring higher performance parameters and requiring steady-state plasma discharges with pulse durations extending to thousands of seconds or longer. To meet these stringent control requirements, a new plasma control system(PCS), named Lingshu, has been independently developed.Similar to existing fusion control frameworks such as MARTe[2], DCS[3][4], and RTF[5][6], Lingshu adopts a component-based architecture characterized by autonomous operation, low coupling, and high performance. Within the framework, control functionalities are encapsulated as Function Blocks (FBs) and deployed inside software components to implement signal acquisition, state diagnosis, plasma control, and command output. A set of supporting services, including workflow scheduling, parameter management, and data archiving, provides the underlying infrastructure for system-wide coordination and reliable operation.The system has been successfully applied to more than 9,000 discharges on the EAST tokamak. Its control components can operate with execution periods as short as 50 s, and simulation results indicate that the framework is capable of supporting steady-state long-pulse operation exceeding 25

A data-driven model for stable long-horizon autoregressive prediction of plasma current and control-oriented boundary evolution in EAST

Minglong Wang, Chenguang Wan, Yuehang Wang, Jia Huang, Zhi Yu, Jingjing Lu, Xiaojuan Liu, Zhisong Qu, Weidong Chen, Teng Wang, et al.

Accurate long-horizon prediction of tokamak plasma current, position, and boundary evolution is essential for magnetic control, rapid controller development, and reinforcement-learning-based optimization. However, high-fidelity physics-based simulators are often computationally prohibitive when large numbers of simulations are required for controller tuning and large-scale optimization. In this work, we develop a fast, data-driven model for the Experimental Advanced Superconducting Tokamak (EAST) to predict plasma current, position, and shape evolution over horizons of up to 1 s, with an inference time of 0.1 s on a single NVIDIA H800 GPU. Evaluated on a temporally separated test set comprising 5907 discharges, the model demonstrates strong generalization to evolving operating conditions and maintains stable agreement with experimental measurements during long-horizon autoregressive rollouts. The mean geometric error of the plasma centroid and the control-oriented boundary representation is 2.4 cm, indicating reliable centimeter-level accuracy for control-oriented prediction of plasma position and shape evolution. The proposed framework provides a practical, high-throughput alternative to first-principles simulators, enabling efficient control algorithm prototyping, large-scale scenario exploration, and future data-driven optimization for tokamak plasma control. This study focuses on the post-shaping phase of EAST discharges, where sustained plasma current and control-oriented boundary regulation are most critical.

Horizon-Aware Early Event Prediction for Tokamak Disruption Alarms

Takeshi Koshizuka, Takaharu Yaguchi

Reliable disruption prediction is essential for the safe operation of future tokamaks. Existing full-distribution survival methods model the complete residual time-to-disruption distribution, whereas operational decisions primarily depend on disruption risk within a finite prediction horizon. This mismatch motivates introducing Early Event Prediction (EEP) objectives into survival-based disruption prediction. We take Deep Survival Machines (DSM) as the full-distribution baseline and propose applying two established EEP methods to tokamak disruption prediction: Temporal Label Smoothing (TLS), which directly predicts disruption probability within a finite horizon, and survTLS, which additionally models the event-time distribution within that horizon. Using a common causal encoder, we compare these methods on DIII-D, Alcator C-Mod, and EAST. We distinguish threshold-free deadline ranking from validation-selected fixed-policy alarm performance and evaluate prediction horizons and encoder architectures. TLS achieves the best mean alarm performance on DIII-D and EAST, whereas all methods perform poorly on Alcator C-Mod. survTLS does not consistently outperform DSM, suggesting that directly learning horizon-level event probability is more effective than modeling detailed within-horizon event-time distributions in the present setting. Finally, the selected prediction horizons and encoder-ablation results vary across devices, reflecting differences in disruption characteristics.

VEQDB: A Compact and Reconstructible Multi-Device Tokamak Equilibrium Database

Huasheng Xie, Ruohan Zhang, Xingyu Li, Feng Zhang, Zhengxiong Wang

Tokamak equilibria are commonly exchanged as gridded G-EQDSK files whose conventions, resolutions, and machine-specific formats impede cross-device comparisons and data-driven modeling. Here, we present VEQDB, an open, compact, and reconstructible fixed-boundary equilibrium database built on continuous MXH--Chebyshev geometry and independent physical-profile roots. By decoupling authoritative equilibrium physics from rectangular meshes, VEQDB enables continuous evaluation and metric differentiation at arbitrary application-demanded resolutions. Backed by an automated numerical validation pipeline, VEQDB is structured as an extensible repository for ongoing community expansion. Its inaugural release provides 13,291 accepted equilibria across 267 conventional and spherical tokamaks, encompassing parameter-sampled Grad--Shafranov solutions, G-EQDSK projections spanning EAST, MAST-U, and ITER scales, and controlled variation families with explicit provenance. Benchmark projections reproduce normalized flux maps with RMS errors between $1.09 \times 10^{-3}$ and $1.45 \times 10^{-3}$, while compact JSON representations achieve an 89--96-fold size reduction relative to standard $129 \times 129$ G-EQDSK files. The complete initial release occupies 41~MB in raw JSON and 18~MB in compressed archives, and all records successfully passed independent reload and evaluation tests. VEQDB establishes an extensible, provenance-preserving foundation for equilibrium studies, reduced-order surrogate modeling, and cross-machine workflows.

Effects of edge-localized mode perturbations and their synergy with tearing modes on fast-ion transport in EAST plasmas: a particle-tracing simulation

Yue Zhang, Mao Li, Feng Wang, Huayi Chang, Pengyun Zhang, Jizhong Sun

Fast-ion transport in EAST plasmas is investigated using a particle-tracing method, focusing on the effects of edge-localized mode (ELM) perturbations and their synergy with tearing-mode (TM) perturbations. The ELM perturbations are obtained from previously reported nonlinear BOUT++ simulations based on EAST experiments and mapped onto the triangular equilibrium mesh used in the Particle orbit Tracing Code, with the TM perturbations prescribed analytically. When only ELM perturbations are considered, the fast-ion loss fraction increases slightly, corresponding to an additional 0.7% of the total fast ions over the simulation duration. The initial positions of the lost ions associated with this increase are concentrated primarily near the plasma boundary around the inner and outer midplanes, while their toroidal LCFS-crossing distribution exhibits a strongly nonuniform pattern. The simulations also show that the parallel component of the ELM-induced electric field can significantly accelerate fast ions, supporting the view that this component contributes to the fast-ion acceleration experimentally observed during ELM events. For coexisting TM and ELM perturbations, the threshold for TM-induced orbit stochasticity remains nearly unchanged, whereas their synergistic effects depend on the TM amplitude. When the TM amplitude is relatively small, the ELM perturbations cause only a modest increase in fast-ion losses, but substantially enhance the local LCFS-crossing density, particularly when the TM amplitude is close to the stochastic threshold. In contrast, when the TM amplitude becomes sufficiently large, the small drift-island chains generated by the TM-ELM synergy, which can already form below the stochastic threshold, reduce the edge orbit stochasticity sufficiently to more than offset the ELM-induced loss enhancement, yielding a net reduction in the overall loss fraction. Although the quantitative loss fractions obtained here could not be regarded as direct predictions for the actual EAST discharge, the physical mechanisms provide useful insight into the characteristic fast-ion response to these perturbations.

Fast electron temperature inference in EAST based on tungsten spectroscopy and convolutional neural networks

Huajian Ji, Zichao Lin, Yongcai Shen, Bo Lyu, HongMing Zhang, Yang Yang, Shihan Huang, Bo Sun

Tungsten (W) spectral line emission from highly charged ions provides an important diagnostic for electron temperature in high-temperature fusion plasmas. In particular, W45+ spectral line intensity ratios exhibit strong sensitivity to electron temperature. However, conventional spectroscopic methods for electron temperature diagnostics typically rely on iterative spectral fitting or collisional-radiative modeling, which can be computationally expensive and limit their applicability for fast analysis. In this work, a new fast electron temperature diagnostic scheme is developed based on W45+ spectral line intensity ratios measured by the X-ray crystal spectrometer (XCS) on EAST. The temperature dependence of the selected line ratios is validated using Flexible Atomic Code (FAC) and FLYCHK calculations, confirming their suitability for temperature diagnostics. A convolutional neural network (CNN) is employed to establish a direct mapping from spectral features to electron temperature, enabling fast inference without iterative computation. The proposed method is validated against electron cyclotron emission (ECE) measurements, showing good agreement with a coefficient of determination of R2 ≈ 0.92 in the temporal evolution. Compared with conventional approaches, the proposed method significantly reduces computational cost while maintaining high fidelity. These results demonstrate that W45+ spectral line ratios provide an effective and efficient observable for electron temperature diagnostics, with strong potential for near-real-time applications in fusion plasma diagnostics.

Automatic Detection of Edge Coherent Modes in EAST Using a Modified YOLO Model

Chengrong Lu, Ying Liu, Shengdi Liao, Yang Ye, Xingli Wang, Zongxiao Guo

Experiment-free disruption prediction for new devices enabled by synthetic diagnostic data augmentation

Zhiqiang Liu, Fengming Xue, Shiwei Xue, Bihao Guo, Dalong Chen, Wei Zheng, Ping Zhu

Deep learning based approaches have shown great promise in cross-device disruption prediction for tokamaks, however, the robustness of these models heavily relies on massive amounts of training data. For the upcoming ITER, to ensure the safety of the first plasma and subsequent operations, experimental data should be entirely unavailable initially, and disruptive discharges should be strictly avoided thereafter. This extreme data scarcity inherently conflicts with the data-intensive nature of deep learning algorithms. To address this challenge, we utilize synthetic diagnostic signals from the target device to supplement the experimental data from existing devices for the zero-shot disruption prediction on a new device. The detailed implementation pipeline of this scheme is presented. For experimental validation, a predictive model trained on data from the EAST tokamak is deployed for a zero-shot cross-device experiment on the J-TEXT tokamak. A synthetic diagnostic framework, configured with the diagnostic parameters of the target device, is developed to process NIMROD magnetohydrodynamic (MHD) simulation data based on the target device's magnetic configuration, thereby achieving effective data augmentation. Ultimately, the results demonstrate that by integrating the target device's synthetic diagnostic data with Fourier Domain Adaptation, the zero-shot accurate early warning rate of the model on 1,596 J-TEXT discharges is improved from 50% to 57%, while exhibiting enhanced predictive robustness.

Plasma-Chemical Degradation of Molybdenum and Tantalum as First Mirror Materials for the EAST Tokamak

Viktor Afanasyev, Lidiia Lobanova, Maxim Semenov-Shefov, Shouxin Wang, Ling Zhang, Haiqing Liu

Effect of pedestal current on the density window for ELM suppression using n = 4 RMP in EAST

Xuemin Wu, Youwen Sun, Qun Ma, Shuai Gu, Manni Jia, Yueqiang Liu, Yifeng Wang, Cheng Ye, Pengcheng Xie, Alberto Loarte, et al.

Existence of operational window in both edge safety factor and line averaged plasma density for suppression of ELMs using n=4 Resonant Magnetic Perturbations in low input torque plasmas has been observed in EAST experiment, in which q95 and plasma normalized beta (βN) close to that required in ITER high-Q operation. Here, n is toroidal mode number of the magnetic perturbation. In contrast to previous reports from other tokamaks, there is not only an upper density limit but also a lower one for accessing ELM suppression. Modelling results using the MARS-F code show that the RMP with linear plasma response has a peak at an intermediate density and decays as the density increases or decreases, which results in a minimal RMP field penetration threshold at the intermediate density. In this experiment, the observed lower density limit operationally manifests a sensitivity of the q-profile: different densities alter the edge current profile, which change the alignment of the eigenmode structure with the RMP coil configuration, causing a reduction of the resonant field in both low- and high-density cases, and hence making field penetration more difficult. The modelled window of the strongest resonant plasma response in terms of [⟨ne⟩, q95] agrees well with the observed ELM suppression in EAST. Peeling-ballooning modes stability analysis using the ELITE code shows that plasmas gradually approach peeling instability boundary caused by increase of edge bootstrap current as the plasma density decreases, which is consistent with the observation that ELMs come back again in lower density plasmas for fixed q95. These results indicate that linear modelling with full toroidal geometry can well predict the optimized RMP configuration for ELM suppression and reveal the important roles of pedestal plasma current, which need to be carefully considered in the application of high n RMPs for ELM suppression in future ITER.

A composite impurity-pressure index for assessing startup readiness during plasma recovery on EAST

Shuqi Yang, Yaowei Yu, Tao Zhang, Xiang Zhu, G Z Zuo, Xiang Gao

Reliable restart after vessel venting is controlled by the burn-through power balance: the plasma must ionize the residual neutrals, dissociate molecules and raise the electron temperature before ionization, charge-exchange and impurity-radiation losses exhaust the available ohmic or auxiliary power. In practice, however, operators usually do not know whether the wall has recovered until a shot has already been attempted. Six EAST plasma-recovery campaigns from 2023 to 2025 are analysed, and a pre-shot impurity-pressure index is introduced to combine the neutral pressure measured 2–3 s before breakdown with a weighted residual-gas-analyser proxy for nitrogen- and oxygen-bearing species. In this dataset, the index organises discharge duration, stable-shot probability and the loop voltage during the first 50–200 ms more clearly than either pressure or composition alone. Under the standard EAST startup condition used here, stable discharges become much more likely when the index falls to around 1 × 10-7 Pa. In the second 2025 campaign, the shot-by-shot evolution also shows threshold-like burn-through behaviour: the maximum line-averaged electron density in 0–0.2 s stays low at large index values, overshoots in a transition interval, and then settles to a more stable level as the wall recovers. Across campaigns, post-recovery values cluster much more tightly than the recovery paths themselves, indicating that the index characterizes the startup-ready wall state rather than a particular conditioning route. The metric therefore provides a practical pre-shot indicator of whether EAST has recovered sufficiently for reproducible stable startup.

Helium‑3 minority heating with ion cyclotron range of frequencies (ICRF) in the experimental advanced superconducting tokamak

Yongxin Zhu, Wei Zhang, Yevgen Kazakov, Jinhua Wu, Paola Mantica, Gabriele Cassella, Tao Jin, X. J. Zhang, Lunan Liu, Hua Yang, et al.

During the 2025 campaign, helium-3 ( 3 He) minority heating with waves in the ion cyclotron range of frequencies (ICRF) was investigated for the first time on the Experimental Advanced Superconducting Tokamak (EAST). With the lowest available ICRF frequency of f IC = 27 MHz, experiments were conducted at a high toroidal magnetic field of B t = 2.8 T and plasma current I p = 450 kA. To optimize 3 He minority heating, the variation of 3 He concentration was systematically explored. Real-time feedback control of the 3 He concentration was successfully implemented through spectroscopic measurement and closed-loop regulation of the 3 He gas injection, demonstrating the feasibility of the control system functions. The ICRF heating efficiency reached a maximum at a minority concentration of ∼8-9%, with the core electron temperature increasing from approximately 5.0 to 7.0 keV and the ion temperature from approximately 1.3 to 1.9 keV under 2.9 MW of ICRF power. These results are in good agreement with simulations from the two-dimensional full-wave code TORIC. Experiments further indicate that higher plasma density enhances 3He heating efficiency. We also briefly discuss the strategy for future 3 He ICRF experiments on EAST.

Numerical studies of mode coupling induced by neoclassical toroidal viscous torque in error field penetration on EAST

Cheng Ye, Youwen Sun, Hui-Hui Wang, Yueqiang Liu, Pengcheng Xie, Jian Xu, Hui Sheng, Xin-Jian Wang, T Y Xia

Mode coupling induced by neoclassical toroidal viscous (NTV) torque governs error field penetration in toroidal plasmas , causing the $2/1$ penetration threshold to deviate from linear response prediction with respect to the RMP upper-lower coil phasing ($\Delta\phi_{UL}$). This behavior originates from nonlinear modulation of toroidal momentum transport by non-resonant three-dimensional magnetic field components. In this work, the characteristics of such mode coupling is systematically investigated using the MARS-Q code. Two aspects are examined. First, the dependence of mode coupling strength on key plasma parameters is analyzed. A larger momentum diffusivity ($\chi_M$) is found to strongly enhance mode coupling and invalidate linear response criteria, whereas at low $\chi_M$ linear prediction remains applicable. A NTV torque weighting factor based on linear response is identified as a qualitative indicator of mode coupling in the nonlinear field penetration. In addition, reduced resistivity ($\eta$), higher normalized beta ($\beta_N$), and larger inverse aspect ratio ($\epsilon$) all strengthen mode coupling. Second, a critical momentum diffusivity, $\chi_{M,crit}$, is introduced to characterize the impact of boundary-induced mode coupling on the $2/1$ penetration spectrum, with smaller value indicating stronger impact. The results show that $\chi_{M,crit}$ increases with initial plasma rotation, but decreases for rotation profile with reduced flow shear. Moreover, increasing $\eta$, as well as higher $\beta_N$ and $\epsilon$ also reduce $\chi_{M,crit}$, albeit through different physical mechanisms. Across broad range of numerical scenarios, these results highlight the essential role of mode coupling in error field control for future fusion devices.

Dimensionless analysis of H-mode plasmas for energy confinement and transport on EAST tokamak

Can Su, Shouxin Wang, Haiqing Liu, Xiao Lan Zou, Shiying Su, Zhuoyang Chen, Zheng Sun, Chen Cheng, Zichao Lin, Guangle Lin, et al.

This work presents an investigation of dimensionless parameter scaling laws for H-mode energy confinement and transport on EAST tokamak, through dedicated experiments designed to isolate the individual effects of normalized plasma pressure ( β ) and collisionality ( ν * ). The β scaling experiment revealed a weak dependence of energy confinement time on β, resulting in a scaling exponent of α β ≈ 0.24±0.20. This weak positive dependence, consistent with experiment observations and linear simulations, reflects an electrostatic turbulence dominated confinement regime in EAST H-mode plasmas. In contrast, a strong, negative dependence of confinement time on ν * was obtained, following the scaling B τ E ∝ ν * -0.71±0.32, indicating confinement improvement with decreasing ν * . Local transport analysis indicates that this strong ν* dependence is primarily associated with electron heat transport, while ion heat transport exhibits a comparatively weak sensitivity. While the linear instability spectrum is dominated by ion temperature gradient (ITG) modes across the explored parameter range, ν * strongly regulates turbulence characteristics relevant to electron heat transport, leading to a pronounced ν* dependence of global confinement. These distinct behaviors of electron and ion heat transport are associated with different physical mechanisms governing the confinement scaling with dimensionless parameters.The present results suggest that the confinement scaling exponents observed on EAST are closely tied to the turbulence regimes accessed in the dedicated β and ν* scans, providing physical insight into the differences with respect to global multi-machine scaling trends.

First wall erosion induced by charge-exchange neutrals on EAST

Rui Ding, Jin Guo, Lei Mu, Guoliang Xu, Yaowei Yu, Yuming Liu, Rong Yan, Hai Xie, Dahuan Zhu, Junling Chen, et al.

Charge-exchange neutrals (CXNs), in particular of hydrogen isotopes deuterium and tritium, are expected to contribute notably to first wall erosion in future fusion reactors. To understand the CXN-induced first wall erosion under different discharge conditions in deuterium, dedicated experiments with a set of new diagnostics have been performed on EAST. Measurements of CXN energy spectrum by the low-energy neutral particle analyzer (LENPA) shows that the integrated CXN flux at the first wall positively correlated with the heating power and line-averaged electron density (n_"e" ), and increased by more than one magnitude from ohmic to high power discharges in the database. Deeper plasma fueling by supersonic molecular beam injection (SMBI) leads to a lower edge neutral pressure and thereby a ~50% lower CXN flux. The CXN flux in the intra-ELM phase is ~2 times higher than that in the inter-ELM phase. Measurements of material erosion rate by the quartz crystal microbalance (QMB) show that higher heating power can lead to stronger material erosion by CXNs. The erosion rate increases with n_"e" at first due to the higher CXN flux and then saturates due to the lower incident energy. The 3D-GAPS code is applied to model the CXN-induced erosion based on the LENPA-measured CXN energy spectrums, which shows good agreement with post-mortem analysis of exposed samples and QMB measurements.

Experimental observation of neoclassical tearing mode stabilization by ICRF drive in EAST

Hua Yang, Wei Zhang, Lunan Liu, Pengjun Sun, tao JIN, Hui-Hui Wang, Liqing Xu, Zhengshuyan Wang, Tonghui Shi, Hailin Zhao, et al.

Neoclassical tearing modes (NTMs) in high-beta plasmas can degrade confinement and trigger disruptions. Experiments were conducted on EAST to investigate the effects of ion cyclotron range of frequency (ICRF) heating on NTMs through controlled variation of the power deposition location and fast-ion distribution. Using hydrogen minority heating, on-axis and off-axis ICRF heating scenarios were achieved by varying the toroidal magnetic field, together with additional ICRF power modulation. The results show that on-axis ICRF heating effectively suppresses the m/n = 3/2 tearing mode, while off-axis heating tends to enhance the m/n = 4/3 mode. On-axis heating also improves plasma confinement and increases the neutron yield. TROIC-TRANSP simulations confirm the distinct power deposition locations, while ASCOT calculations indicate that the fast-ion energy reaches up to 800 keV during on-axis heating, which is substantially higher than that in the off-axis cases. A modified Rutherford equation incorporating fast-ion effects suggests that the fast-ion-driven uncompensated cross-field current term is responsible for the observed NTM behaviors. These results demonstrate that controlling the ICRF resonance position is a feasible approach for NTM suppression and improved plasma performance.

Analysis of neutron emission during NBI–ICRF synergistic heating in EAST high neutron rate high β p discharges

Andong Xu, Mingyuan Xu, Yunhe Li, Tao Yu, Jiayi Zhang, Yubo Zhang, Yongqiang Zhang, Chenyu Pan, Baolong Hao, Pan Li, et al.

This paper reports the analysis of neutron emission characteristics in high poloidal beta (β p ) discharges on the EAST tokamak, where a record fusion neutron rate of S n = 3.9×10 14 s -1 was achieved with β p ∼2.8, β N ∼2.2, and H 98,y2 ∼1.3. Statistical analysis reveals that while ion cyclotron range of frequencies (ICRF) heating significantly boosts both the neutron rate and plasma stored energy, the neutron rate scales sub-linearly with neutral beam injection (NBI) power (S n ∝P NBI 0.92 ). Interpretive TRANSP simulations demonstrate that the NBI--ICRF synergistic effect directly contributes approximately 30% to the total neutron rate through the formation of a high-energy fast-ion tail. However, this enhancement is partially offset by NBI-induced profile degradation, including fuel dilution, impurity accumulation, and core electron temperature reduction. The generation of the fast-ion tail in velocity space is validated by multi-sightline neutron emission spectroscopy. Furthermore, orbit topology analysis using the ORBIT code reveals that the synergistic effect drives suprathermal fast ions into smaller orbits, such as stagnation orbits, leading to a spatial redistribution of fast ions and a consequent peaking of the neutron emissivity profile. These findings provide critical insights into the complex interplay between auxiliary heating, fast-ion behavior, and neutron emission, offering valuable insights for achieving higher beam--thermal fusion rates in future deuterium--deuterium and deuterium--tritium experiments.

The role of data quality and alignment in cross-tokamak disruption prediction

Chengshuo Shen, Wei Zheng, Fengming Xue, Xinkun Ai, Bihao Guo, Dalong Chen, Zhongyong Chen, Yong Hua Ding

Reliable disruption prediction across tokamaks is needed for next generation devices such as ITER, SPARC, and BEST, where disruptive target-machine data will be scarce by design. For a transferred predictor, target performance is limited by both the source-domain error and the mismatch between the source and target feature distributions. We use this distinction to revisit our previous J-TEXT to EAST study and to add three components for few-shot and zero-shot operation. An upgraded physics-guided feature extraction (PGFE-U) reduces geometry-dependent differences in Mirnov, soft x-ray and absolute extreme ultraviolet array features. A floating labelling strategy (FLS) replaces a fixed pre-disruption window with shot-dependent precursor onset labels. An estimation of the feature distribution (EFD) supplies target-machine z-score statistics from outside the classifier training set. With the supervised CORAL (S-CORAL) backbone, the area under the receiver operating characteristic curve (AUC) on the same EAST test set reaches 0.957 in a few-shot setting using only 10 disruptive EAST discharges, whereas the previous work required 110 EAST discharges to reach 0.890, and the partial AUC over false positive rates up to 10% (pAUC) rises from 0.428 to 0.858. The zero-shot AUC reaches 0.892 with no EAST discharge used for classifier training, against 0.592 for the same model normalised with J-TEXT statistics, and the zero-shot pAUC is 0.619 against 0.084. The EFD statistics are estimated from real EAST shots; a per-feature Monte-Carlo scan shows that the sensitivity is concentrated in a few global discharge parameters, the quantities most reliably estimated at the design stage. Shapley additive explanation (SHAP) based attribution shows that both transferred models preserve most of the qualitative feature-to-disruption trends of the abundant-data baseline. These results indicate that improving feature quality, labels and normalization statistics can recover a large fraction of abundant-data performance when target-machine shots are limited.