We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Plasma Physics

Title: Measuring the Electron Temperature and Identifying Plasma Detachment using Machine Learning and Spectroscopy

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 $10\times$ 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.
Comments: Submitted to Review of Scientific Instruments
Subjects: Plasma Physics (physics.plasm-ph); Instrumentation and Detectors (physics.ins-det)
Report number: LLNL-JRNL-815130
Cite as: arXiv:2010.11244 [physics.plasm-ph]
  (or arXiv:2010.11244v1 [physics.plasm-ph] for this version)

Submission history

From: Cameron Samuell [view email]
[v1] Wed, 21 Oct 2020 18:59:31 GMT (4813kb,D)

Link back to: arXiv, form interface, contact.