No Arabic abstract
A machine learning approach has been implemented to measure the electron temperature directly from the emission spectra of a tokamak plasma. This approach utilized a neural network (NN) trained on a dataset of 1865 time slices from operation of the DIII-D tokamak using extreme ultraviolet / vacuum ultraviolet (EUV/VUV) emission spectroscopy matched with high-accuracy divertor Thomson scattering measurements of the electron temperature, $T_e$. This NN is shown to be particularly good at predicting $T_e$ at low temperatures ($T_e < 10$ eV) where the NN demonstrated a mean average error of less than 1 eV. Trained to detect plasma detachment in the tokamak divertor, a NN classifier was able to correctly identify detached states ($T_e<5$ eV) with a 99% accuracy (F$_1$ score of 0.96) at an acquisition rate $10times$ faster than the Thomson scattering measurement. The performance of the model is understood by examining a set of 4800 theoretical spectra generated using collisional radiative modeling that was also used to predict the performance of a low-cost spectrometer viewing nitrogen emission in the visible wavelengths. These results provide a proof-of-principle that low-cost spectrometers leveraged with machine learning can be used both to boost the performance of more expensive diagnostics on fusion devices, and be used independently as a fast and accurate $T_e$ measurement and detachment classifier.
We show that the charge accumulated by a dielectric plasma-facing solid can be measured by infrared spectroscopy. The approach utilizes a stack of materials supporting a surface plasmon resonance in the infrared. For frequencies near the Berreman resonance of the layer facing the plasma the reflectivity dip--measured from the back of the stack, not in contact with the plasma--depends strongly on the angle of incidence making it an ideal sensor for the changes of the layers dielectric function due to the polarizability of the trapped surplus charges. The charge-induced shifts of the dip, both as a function of the angle and the frequency of the incident infrared light, are large enough to be measurable by attenuated total reflection setups.
A simple table-size ECR plasma generator operates in the ATOMKI without axial magnetic trap and without any particle extraction tool. Radial plasma confinement is ensured by a NdFeB hexapole. The table-top ECR is a simplified version of the 14 GHz ATOMKI-ECRIS. Plasma diagnostics experiments are planned to be performed at this device before installing the measurement setting at the big ECRIS. Recently, the plasma generator has been operated in pulsed RF mode in order to investigate the time evolution of the ECR plasma in two different ways. (1) The visible light radiation emitted by the plasma was investigated by the frames of a fast camera images with 1 ms temporal resolution. Since the visible light photographs are in strong correlation with the two-dimensional spatial distribution of the cold electron components of the plasma it can be important to understand better the transient processes just after the breakdown and just after the glow. (2) The time-resolved ion current on a specially shaped electrode was measured simultaneously in order to compare it with the visible light photographs. The response of the plasma was detected by changing some external setting parameters (gas pressure and microwave power) and was described in this paper.
The GlueX forward calorimeter is an array of 2800 lead glass modules that was constructed to detect photons produced in the decays of hadrons. A background to this process originates from hadronic interactions in the calorimeter, which, in some instances, can be difficult to distinguish from low energy photon interactions. Machine learning techniques were applied to the classification of particle interactions in the GlueX forward calorimeter. The algorithms were trained on data using decays of the $omega$ meson, which contain both true photons and charged particles that interact with the calorimeter. Algorithms were evaluated on efficiency, rate of false positives, run time, and implementation complexity. An algorithm that utilizes a multi-layer perceptron neural net was deployed in the GlueX software stack and provides a signal efficiency of 85% with a background rejection of 60% for an inclusive $pi^0$ data sample for an intermediate quality constraint.
Single-shot absorption measurements have been performed using the multi-keV X-rays generated by a laser wakefield accelerator. A 200 TW laser was used to drive a laser wakefield accelerator in a mode which produced broadband electron beams with a maximum energy above 1 GeV and a broad divergence of $approx15$ miliradians FWHM. Betatron oscillations of these electrons generated $1.2pm0.2times10^6$ photons/eV in the 5 keV region, with a signal-to-noise ratio of approximately 300:1. This was sufficient to allow high-resolution XANES measurements at the K-edge of a titanium sample in a single shot. We demonstrate that this source is capable of single-shot, simultaneous measurements of both the electron and ion distributions in matter heated to eV temperatures by comparison with DFT simulations. The unique combination of a high-flux, large bandwidth, few femtosecond duration X-ray pulse synchronised to a high-power laser will enable key advances in the study of ultra-fast energetic processes such as electron-ion equilibration.
In this work we provide experimental insights into the impact of plasma-molecule interactions on the target ion flux decrease during divertor detachment achieved through a core density ramp in the TCV tokamak. Our improved analysis of the hydrogen Balmer series shows that plasma-molecule processes are strongly contributing to the Balmer series intensities and substantially alter the divertor detachment particle balance. We find that Molecular Activated Recombination (MAR) ion sinks from $H_2^+$ and/or $H^-$ are a factor $sim$ 5 larger than Electron-Ion Recombination (EIR) and are a significant contributor to the observed reduction in the outer divertor ion target flux. Molecular Activated Ionisation (MAI) may also be significant during detachment. Plasma-molecule interactions enhance the Balmer line series emission strongly near the target as detachment proceeds. This indicates enhancements of the Lyman series, potentially affecting power balance in the divertor. As those enhancements vary spatially in the divertor and are different for different transitions, they are expected to result in a separation of the $Lybeta$ and $Lyalpha$ emission regions. This may have implications for the treatment and diagnosis of divertor opacity. The demonstrated enhancement of the Balmer series through plasma-molecule processes potentially poses a challenge to using the Balmer series for understanding and diagnosing detachment based only on atom-plasma processes.