Advanced vacuum technologies, including pumping, fueling and wall conditioning, have been successfully developed or upgraded to efficiently control the fuel and impurity particles to extend the plasma pulse duration in the experimental advanced superconducting tokamak (EAST). To improve the particle exhaust rate cryopumps with a 60% increase in pumping speed and similar to 2 times increase in saturation capacity have been developed, and molecular pumps with a similar to 30% increase in pumping speed have been upgraded. In order to monitor the molecular pump status while avoiding bearing faults and overload accidents, a fault detection system has been built which can offer an early warning to avoid more losses within the fusion device. A series of fueling technologies have been developed including gas injection system, supersonic molecular beam injector, pellet injector (PI), massive gas injector and shattered pellet injector, installed at the midplane and divertor positions at different ports to improve fueling uniformity and efficiency. Meanwhile, routine wall conditioning such as electric and hot N-2 baking, ion cyclotron wall conditioning and glow discharge cleaning have been successfully developed to remove the impurity particles from the inner component and materials. The low Z material wall coating and real-time powder injection during plasma discharge are also designed and applied to further improve particle control capability. By using these advanced vacuum related technologies, good vacuum (<2 x 10(-6)Pa) and wall conditions are realized, and the fuel and impurity particles can be effectively and stably controlled, which promotes the achievement of the record plasma of similar to 1056 s pulse duration with the line-averaged electron density of 1.8 x 10(19) m-3 on EAST. They provide a very important reference for vacuum system design and operation for future fusion devices.

Remote sensing image semantic segmentation has extensive applications in land resource planning and smart cities. Due to the problems of unclear boundary segmentation and insufficient Semantic information of small targets in high-resolution remote sensing images, an improved network TU net based on U-net++ is proposed. Secondly, the attention aggregation module of the base Transformer is introduced to capture global contextual information, replacing the original multi-level skip connections of U-net++. A cross-window interaction module is designed, which significantly reduces computational complexity and achieves a lightweight model. Finally, a dynamic feature fusion block is designed at the end of the decoder to obtain multi-class and multi-scale Semantic information and enhance the final segmentation effect. TU-net conducted experiments on two datasets, where OA, mIoU, and mF1 scores were higher than mainstream models. The IoU and F1 scores of small-sized target cars in the Vaihingen dataset were 0.896 and 0.962, respectively, which were 5% and 15.8% higher than the suboptimal model; The IoU and F1 scores of the trees in the Potsdam dataset are 0.913 and 0.936, respectively, which are 6.3% and 4.3% higher than the suboptimal model. The experimental results show that the model can more accurately segment small-sized targets and target boundaries.

Recent research has been conducted on the removal of retained fuel particles from the first wall surfaces in the EAST superconducting tokamak, by using direct-current Glow Discharge Cleaning (GDC) under strong magnetic field. The findings from these experiments reveal that GDC can operate in a toroidal field range of 0-2.5 T and achieve high fuel removal rate despite the plasma is strongly confined by the magnetic field. The cleaning process primarily targets the side portion of the limiter adjacent to the GDC anodes on the low field side via thermal desorption to access areas that are typically inaccessible to other plasma types. Additionally, pulsed GDC is less effective than continuous GDC operation under intense magnetic field conditions. The efficiency of pulsed GDC mainly depends on the GDC duty cycle, which further suggests that thermal desorption is the predominant cleaning mechanism in the strong magnetic environment. The integration of Ion Cyclotron Wall Conditioning (ICWC) with GDC in a pulsed mode yields an efficiency increase of roughly 35% over ICWC alone. Furthermore, the new integration cleaning mode can effectively extend the cleaning area. A potential synergy with a variety of discharge cleaning techniques could further extend the cleaning area and improve the efficiency of tritium removal. These insights serve as crucial benchmarks for the removal of retained tritium in the presence of strong magnetic fields in future fusion reactors.

In the 2018 EAST (Experimental Advanced Superconducting Tokamak) experiments, type-I ELMs (edge localized modes) were mitigated and suppressed by Ne-SMBI (supersonic molecular beam injection) injection during H-mode discharges. No energy confinement degradation was observed during ELM suppression, whereas approximately 10% degradation was observed during ELM mitigation. The observation of broadband MHD (magnetohydrodynamic) turbulence during ELM suppression indicates it may be the primary particle transport channel. The electron density profile in the pedestal top increased and the edge radiation increased with Ne-SMBI injection, which could reduce heat fluxes reaching the divertor. Pedestal profile changes appear adequate to trigger small transport events and mitigate large ELMs. In the 2022 EAST experimental campaign, stable 2 s ELM suppression was achieved by adjusting the Ne-SMBI injection rate. High-frequency coherent modes around 250 kHz, possibly corresponding to micro-tearing modes (MTM), appeared during ELM suppression. In addition to impurity SMBI injection, fuel particle D2-SMBI injection was also applied for ELM mitigation. Experimental results show D2-SMBI has small impact on the edge, increasing ELM frequency without significant ELM mitigation compared to impurity SMBI.

Since the last IAEA-FEC in 2021, significant progress on the development of long pulse steady state scenario and its related key physics and technologies have been achieved, including the reproducible 403 s long-pulse steady-state H-mode plasma with pure radio frequency (RF) power heating. A thousand-second time scale (similar to 1056 s) fully non-inductive plasma with high injected energy up to 1.73 GJ has also been achieved. The EAST operational regime of high beta(P) has been significantly extended (H-98y2 > 1.3, beta(P) similar to 4.0, beta(N) similar to 2.4 and n(e)/n(GW) similar to 1.0) using RF and neutral beam injection (NBI). The full edge localized mode suppression using the n = 4 resonant magnetic perturbations has been achieved in ITER-like standard type-I ELMy H-mode plasmas with q(95) approximate to 3.1 on EAST, extrapolating favorably to the ITER baseline scenario. The sustained large ELM control and stable partial detachment have been achieved with Ne seeding. The underlying physics of plasma-beta effect for error field penetration, where toroidal effect dominates, is disclosed by comparing the results in cylindrical theory and MARS-Q simulation in EAST. Breakdown and plasma initiation at low toroidal electric fields (<0.3 V m(-1)) with EC pre-ionization is developed. A beneficial role on the lower hybrid wave injection to control the tungsten concentration in the NBI discharge is observed for the first time in EAST suggesting a potential way toward steady-state H-mode NBI operation.

Particle exhaust is one of the key factors in controlling fuel recycling and enhancing plasma performance. Non-Evaporable Getter (NEG) pumps, known for their high-temperature resistance and large pumping speed, are potentially suitable for enhancing the particle exhaust capability of the tokamak. Extensive studies were conducted to evaluate their pumping performance in the Vacuum testing system and EAST tokamak. Experiments have shown that NEG pumps stably operate from room temperature to 200°C, displaying a clear positive correlation between pumping speed and temperature. These pumps outperformed traditional cryopumps, particularly at standard D2 pressures of above 0.1 Pa. Moreover, nitrogen passivation has been proven to effectively limit pumping speed and prevent hydrogen embrittlement. Despite undergoing rigorous operational conditions, including a total of 30 hours of plasma, lithium processes, atmospheric exposures and water leakages, the NEG pumps retained 70% of their initial speed. These findings suggest that NEG pumps are compatible with plasma operation environments, providing a promising solution for particle exhaust in future fusion reactors.

Abstract An intelligent control system for Non-Evaporable Getter (NEG) pumps was designed and developed. The main functions of the system include remote monitoring and remote control of key parameters of the NEG power supply, intelligent control of NEG working status according to NEG saturation status and vacuum chamber status, especially intelligent warning and protection of NEG from vacuum leaks and NEG supersaturation. The intelligent control system is proven stable and reliable, which provides a feasible logic architecture and control design for NEG pump application in future fusion devices.

In this paper, experimental and numerical studies are conducted on supersonic molecular beam injection (SMBI) for fusion devices based on steady-state assumption. To explore the characteristics of SMBI from continuous to rarefied flow, the flow information is predicted by solving the Navier-Stokes (NS) equation based CFD method and direct simulation Monte Carlo (DSMC) method. To verify the coupled NS-DSMC method, a flow measurement apparatus platform is designed and constructed. A moveable full-range vacuum gauge with Pitot probe in the vacuum chamber is employed to measure the spatial distribution of the beam pressure signal. The simulation results are converted into Pitot pressure, corrected for rarefaction effect and compared with the experimental data. It is observed that the experimental results are consistent with the simulation results. This study provides technical support for the accurate prediction of SMBI flow characteristics and optimization of the SMBI systems.

EAST implements the double feedback control experiments of the detachment by using two control systems at the same time to explore the new detachment control way. We hire two PID controllers and two gas puffing valves dividedly to inject impurity gases from different locations, increasing the radiated power of the bulk plasma and reducing the divertor electron temperature (Te,t) simultaneously. This control method is applied in the EAST long-pulse H-mode plasma, the controlled variables of the radiative feedback control is the local radiative intensity and the total radiated power respectively. In the experiments, the longest double control duration has been up to 13.5 s, the divertor heat flux is reduced significantly in the control duration, while the particle flux has an overall drop at the end of the control phase. The chosen impurity species in the control system are Neon (Ne) and Argon (Ar): Ne is injected from the upper divertor to raise the core radiated power, and the Ar is injected from the lower divertor to reduce the Te,t beneath 8 eV. However, the Ar injection is not only decreasing the Te,t, but also generates an unfavorable increment of the core radiated power. This increment makes the bulk radiated power is higher than the target value, the radiative feedback control thus is paused for a long time. This mechanism results in the double feedback control back to the sequential implementations of the two feedback controls, which does not accord with the original expection(the both controls run at the same time), and degrades the control precision. In the future, many mitigating ways will be tested to improve the control effect. We will also develop a dynamic response model of the impurity seeding to simulate the time evolution of the divertor heat load with two kinds of impurities. The modeling results will be used to optimize the double feedback control system to gain a better experimental results.

A coherent mode (CM) induced by the supersonic molecular beam injection (SMBI) was observed in Ohmic plasmas of EAST tokamak. The CM existed in the spectra of the electron temperature fluctuation and density fluctuation, but not in the spectrum of Mirnov fluctuation. The radial location of the CM was about 10 cm inside the separatrix (normalized minor radius ρ∼0.75 ). In most cases, the CM has multiple discrete frequencies between 5∼20kHz with an interval of ∼1 kHz. Occasionally, there is only one dominant frequency. The poloidal scale of the CM (poloidal wavenumber kθ∼0.4cm−1 , poloidal mode number m∼20 ) is between that of typical ion turbulence and macroscopic magnetohydrodynamic activities. Statistical analysis results suggest that the onset of the CM requires a high temperature gradient and a relatively low density gradient. Furthermore, the transport analysis shows that the SMBI-caused perturbation propagates inward faster in plasmas with stronger CM, which implies that the CM might be correlated to a high local transport level. However, the impact of the CM on macroscopic plasma parameters including stored energy, energy confinement time, and SMBI-caused mean density fluctuation is too weak to be found in comparison with the random fluctuation and the measurement errors, which suggests that the CM does not have a clear influence on the global confinement.

An improved prediction algorithm of the hidden semi-markov model (HSMM) is proposed to predict the remaining useful life (RUL) of aluminum reduction cells (ARC). First, the degradation process of the ARC is analyzed, and the parameters that could characterize its degradation process are determined. Second, to facilitate the prediction of HSMM, exponential distribution, in which form the dwell time in the traditional HSMM is distributed, is replaced with the Erlang distribution based on the queuing theory. Third, an improved forward recursion algorithm that introduced dwell time has not only simplified the calculation of useful life predictions but also further facilitated HSMM in the prediction of the RUL. Finally, verification is carried out using the actual data of an aluminum electrolytic industry, and the results showed that the improved HSMM performed better in the prediction and is more accurate than the existing prediction method.

In order to overcome the dynamic and large-scale characteristics of the plant-wide processes, this paper proposed a distributed slow feature analysis (SFA) with inter-unit dissimilarity method for process monitoring task. Firstly, to highlight the local dynamic features, the whole process is decomposed into several units according to the prior knowledge. Based on this, SFA monitoring model is built parallelly to handle the dynamic features. Considering the possible information loss caused by the process decomposition, the inter-unit dissimilarity index is carried out to monitor the variations between adjacent units. Finally, the fusion center is conducted by Bayesian inference to combine the results of SFA monitoring models and inter-unit dissimilarity statistics. The effectiveness of the proposed method is tested on the Tennessee Eastman process and an aluminum electrolysis process.

The divertor is the main device in the tokamak that is subjected to heat loads, excessive high heat flux could cause damage to the divertor target. Divertor heat flux control is very important for magnetically confined fusion devices. As a method for divertor heat flux control, the Divertor Fast Particles Injection (DFPI) is successfully developed and implemented on the Experimental Advanced Superconducting Tokamak (EAST). The components of DFPI are a high-pressure gas source (0.6-1 MPa), a pulse solenoid valve, a nozzle and the pipeline etc. Compared with the Supersonic Molecular Beam Injection (SMBI) particle injection from midplane, the particles injected by DFPI are more difficult to enter the plasma core and have less impact on the performance of the core plasma; Compared with the gas puffing (GP) particle injection from the divertor, the particles injected by DFPI can enter the divertor area faster, the delay time of DFPI is about 25 ms. The successful implementation of DFPI provides an important tool for divertor heat flux control on EAST.

针对铝电解槽中氧化铝浓度无法实时测量、传统控制方法过分依赖专家经验且难以精准控制的问题,提出了基于非线性模型预测控制方法(Nonlinear model predictive control,NMPC)的氧化铝浓度精确控制策略.首先,通过铝电解生产机理分析,确定了NMPC内部模型(简称内模)的输入输出变量;其次,采用基于最小二乘支持向量机的非线性Hammerstein系统子空间辨识方法,建立了数据驱动的氧化铝浓度控制的状态空间模型,同时,为了降低模型计算的复杂度,用多项式拟合模型中表示非线性特性的核函数;再次,根据拟合多项式的反函数,设计了非线性控制系统结构和性能指标函数.最后,基于某铝厂实际生产数据验证了所提控制方法的有效性和正确性.

Particle control is one of the key issues for steady-state tokamak operation. The density decay time is used to evaluate the capability of the divertor to exhaust particle with all fueling turn off. Experiments of the EAST divertor's ability to exhaust helium particles were also performed during the EAST helium plasma operation. The experimental results show that the EAST divertor is less capable of exhaust helium particles than deuterium particles. And similar to deuterium, the helium particle exhaust performance is sensitivity on the different strike point location. During 2021 experiment campaign, a new lower tungsten divertor has been developed and installed on EAST, the particle exhaust capacity of the new divertor has also been investigated during this campaign.

Transition from type-III to large-amplitude ELMs induced by neon injection has been observed in the EAST tokamak at overlapping q(95) space between large and small ELMs. With neon injection, pedestal density gradient shows a remarkable increase accompanied by some decrease of pedestal electron temperature, and consequently the pressure gradient increases moderately and edge bootstrap current has minimal change. Further experiment demonstrates that the occurrence of large ELMs after neon injection is highly correlated with the change in edge density. Linear peeling-ballooning stability analysis indicates that the large ELM case is more unstable than the type-III ELM case during the ELM transition. A scan of pedestal density gradient in linear stability analysis shows that the direct destabilizing effect of steep pedestal density gradient on peeling-ballooning instabilities via two-fluid effects could also facilitate the transition to large ELMs. These results could provide more insight into the role of pedestal density gradient on pedestal stability and ELM behavior. (C) 2022 Elsevier B.V. All rights reserved.

With the continuous development of large-scale aluminum reduction cells, the problem of the uniform distribution of alumina concentration in the cell has become more and more serious for the reduction process. In order to achieve the uniform distribution of the alumina concentration, a data-driven distributed subspace predictive control feeding strategy is proposed in this paper. Firstly, the aluminum reduction cell is divided into multiple sub-systems that affect each other according to the position of the feeding port. Based on the subspace method, the prediction model of the whole cell is identified, and the prediction output expression of each sub-system is deduced by decomposition. Secondly, the feeding controller is designed for each aluminum reduction cell subsystem, and the input and output information can be exchanged between each controller through the network. Thirdly, under consideration of the influence of other subsystems, each subsystem solves the Nash-optimal control feeding quantity, so that each subsystem realizes distributed feeding. Finally, the simulation results show that, compared with the traditional control method, the proposed distributed feeding control strategy can significantly improve the problem of the uniform distribution of alumina concentration and improve the current efficiency of the aluminum reduction cell.

With the increasing demand for petroleum resources and environmental issues, new energy electric vehicles are increasingly being used. However, the large number of electric vehicles connected to the grid has brought new challenges to the operation of the grid. Firstly, A novel bidirectional interaction model is established based on modulation theory with nonlinear loads. Then, the electric energy measuring scheme of EVs for V2G is derived under the conditions of distorted power loads. The scheme is composed of fundamental electric energy, fundamental-distorted electric energy, distorted-fundamental electric energy and distorted electric energy. And the characteristics of each electric energy are analyzed. Finally, the correctness of the model and energy measurement method is verified by three simulation cases: the impact signals, the fluctuating signals, and the harmonic signals.

Deuterium recycling properties have been investigated in the EAST superconducting tokamak by evaluating the density decay time after the fueling termination. The density decay time of the latter discharge of the experiment of a day is ∼2.9 s, which is more than three times longer than that of the beginning discharge (∼0.88 s), indicating that a repeat of discharges diminishes the wall conditioning effect by lithium coating. However, the difference in deuterium recycling among three magnetic configurations (lower single null, double null, and upper single null) is unclear. The density decay times are 3.3 s and 4.1 s when the strike point (SP) positions are located on the horizontal and vertical plates of the lower outer divertor, showing that the particle exhaust capability becomes stronger as the SP position approaches the pumping slot. Furthermore, the lower hybrid current drive (LHCD) power enhances deuterium recycling since the density decay times of high (∼2.7 MW) and low (∼0.8 MW) power LHCD discharges are 4.3 s and 1.2 s, respectively. The density decay time in the LHCD discharge linearly increases with an increase in a period of exponential fitting, implying that the recycling coefficient increases during the stop of fueling. However, the density decay time in the ohmic discharge is independent of the fitting time.

A method of local anode effect prediction is proposed for the problem that it is difficult to detect the local anode effect in large aluminum reduction cell in real time. Firstly, a fuzzy classification of local anode effect prediction in terms of fuzziness level is proposed considering various working conditions of anode current in the region. Secondly, a current volatility detection method based on time-sliding window density is designed from the problem of uneven current distribution in the region, and the anode currents in the region are classified and tracked for prediction according to the different current volatility. Thirdly, an improved Gated Recurrent Unit (GRU) neural network structure is proposed to improve the prediction accuracy of fluctuating currents. Finally, simulation experiments are conducted based on actual data, and compared with Long Short-Term Memory (LSTM) and Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM), the proposed method has certain advantages in both prediction time, training time, and the mean absolute error (MAE) and mean square error (MSE), which verifies the effectiveness of the proposed method.