Refine
Year of publication
- 2022 (1)
Document Type
- Article (1)
Language
- English (1) (remove)
Has Fulltext
- yes (1)
Is part of the Bibliography
- no (1)
Keywords
- - (1)
- W7-X (1)
- data-mining (1)
- entropy (1)
- permutation (1)
Institute
Publisher
- IOP Publishing (1)
Detecting changes in plasmas is compulsory for control and the detection of novelties.
Moreover, automated novelty detection allows one to investigate large data sets to substantially
enhance the efficiency of data mining approaches. To this end we introduce permutation entropy
(PE) for the detection of changes in plasmas. PE is an information-theoretic complexity measure
based in fluctuation analysis that quantifies the degree of randomness (resp. disorder,
unpredictability) of the ordering of time series data. This method is computationally fast and
robust against noise, which allows the evaluation of large data sets in an automated procedure.
PE is applied on electron cyclotron emission and soft x-ray measurements in different
Wendelstein 7-X low-iota configuration plasmas. A spontaneous transition to high core-electron
temperature (Te) was detected, as well as a localized low-coherent intermittent oscillation which
ceased when Te increased in the transition. The results are validated with spectrogram analysis
and provide evidence that a complexity measure such as PE is a method to support in-situ
monitoring of plasma parameters and for novelty detection in plasma data. Moreover, the
acceleration in processing time offers implementations of plasma-state-detection that provides
results fast enough to induce control actions even during the experiment.