اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدExperiment data-driven modeling of tokamak discharge in EAST
89
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
A model for tokamak discharge through deep learning has been done on a superconducting long-pulse tokamak (EAST). This model can use the control signals (i.e. Neutral Beam Injection (NBI), Ion Cyclotron Resonance Heating (ICRH), etc) to model normal discharge without the need for doing real experiments. By using the data-driven methodology, we exploit the temporal sequence of control signals for a large set of EAST discharges to develop a deep learning model for modeling discharge diagnostic signals, such as electron density $n_{e}$, store energy $W_{mhd}$ and loop voltage $V_{loop}$. Comparing the similar methodology, we use Machine Learning techniques to develop the data-driven model for discharge modeling rather than disruption prediction. Up to 95% similarity was achieved for $W_{mhd}$. The first try showed promising results for modeling of tokamak discharge by using the data-driven methodology. The data-driven methodology provides an alternative to physical-driven modeling for tokamak discharge modeling.
قيم البحث
اقرأ أيضاً
The optimum scheme for geometric phase measurement in EAST Tokamak is proposed in this paper. The theoretical values of geometric phase for the probe beams of EAST Polarimeter-Interferometer (POINT) system are calculated by path integration in parame
ter space. Meanwhile, the influences of some controllable parameters on geometric phase are evaluated. The feasibility and challenge of distinguishing geometric effect in the POINT signal are also assessed in detail.
Evolution of electron temperature, electron density and its fluctuation with high spatial and temporal resolutions are presented for the cold pulse propagation induced by super-sonic molecular beam injection (SMBI) in ohmic plasmas in the EAST tokama
k. The non-local heat transport occurs for discharges with plasma current $I_p$=450 kA ($q_{95}sim5.55$), and electron density $n_{e0}$ below a critical value of $(1.35pm0.25)times10^{19}~mathrm{m^{-3}}$. In contrary to the response of core electron temperature and electron density (roughly 10 ms after SMBI), the electron density fluctuation in the plasma core increases promptly after SMBI and reaches its maximum around 15 ms after SMBI. The electron density fluctuation in the plasma core begins to decrease before the core electron temperature reaches its maximum (roughly 30 ms). It was also observed that the turbulence perpendicular velocity close to the inversion point of the temperature perturbation changes sign after SMBI.
One metric for comparing confinement properties of different magnetic fusion energy configurations is the linear critical gradient of drift wave modes. The critical gradient scale length determines the ratio of the core to pedestal temperature when a
plasma is limited to marginal stability in the plasma core. The gyrokinetic turbulence code GS2 was used to calculate critical temperature gradients for the linear, collisionless ion temperature gradient (ITG) mode in the National Compact Stellarator Experiment (NCSX) and a prototypical shaped tokamak, based on the profiles of a JET H-mode shot and the stronger shaping of ARIES-AT. While a concern was that the narrow cross section of NCSX at some toroidal locations would result in steep gradients that drive instabilities more easily, it is found that other stabilizing effects of the stellarator configuration offset this so that the normalized critical gradients for NCSX are competitive with or even better than for the tokamak. For the adiabatic ITG mode, NCSX and the tokamak had similar critical gradients, though beyond marginal stability, NCSX had larger growth rates. However, for the kinetic ITG mode, NCSX had a higher critical gradient and lower growth rates until a/L_T is approximately 1.5 times a/L_{T,crit}, when it surpassed the tokamaks. A discussion of the results presented with respect to a/L_T vs. R/L_T is included.
Through a systematically developed theory, we demonstrate that the motion of instanton identified in [Y. Z. Zhang, Z. Y. Liu, T. Xie, S. M. Mahajan, and J. Liu, Physics of Plasmas 24, 122304 (2017)] is highly correlated to the intermittent excitation
and propagation of geodesic acoustic mode (GAM) that are observed in tokamaks. While many numerical simulations have observed the phenomena, it is the first theory that reveals the physical mechanism behind GAM intermittent excitation and propagation. The preceding work is based on the micro-turbulence associated with toroidal ion temperature gradient (ITG) mode, and slab-based phenomenological model of zonal flow. When full toroidal effect are introduced into the system, two branches of zonal flow emerge: the torus-modified low frequency zonal flow (TLFZF), and GAM, necessitating a unified exploration of GAM and TLFZF. Indeed, we observe that the transition (decay) from the caviton to instanton is triggered by a rapid zero-crossing of radial group velocity of drift wave and is found to be strongly correlated with the GAM onset. Many features peculiar to intermittent GAMs, observed in real machines, are thus identified in the numerical experiment. The results will be displayed in figures and in a movie; first for single central rational surface, and then with coupled multiple central rational surfaces. The periodic bursting first shown disappears as being replaced by irregular one, more similar to the intermittent characteristics observed in GAM experiments.
Axisymmetric free-boundary equilibrium reconstructions of tokamak plasmas in the Lithium Tokamak eXperiment (LTX) are performed using the PSI-Tri equilibrium code. Reconstructions in LTX are complicated by the presence of long-lived non-axisymmetric
eddy currents generated by vacuum vessel and first wall structures. To account for this effect, reconstructions are performed with additional toroidal current sources in these conducting regions. The source distributions are fixed poloidally, but their scale is adjusted as part of the full reconstruction. Eddy distributions are computed by toroidally averaging currents, generated by coupling to vacuum field coils, from a simplified 3D filament model of important conducting structures. The full 3D eddy current fields are also used to enable the inclusion of local magnetic field measurements, which have strong 3D eddy current pick-up, as reconstruction constraints. Using this method, equilibrium reconstruction yields good agreement with all available diagnostic signals. An accompanying field perturbation produced by 3D eddy currents on the plasma surface with primarily n=2, m=1 character is also predicted for these equilibria.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
