Implementation of the disruption predictor APODIS in JET real time network using the MARTe framework

J.M. Lopez, J. Vega, D. Alves, S. Dormido-Canto, A. Murari, J.M. Ramirez, R. Felton, M. Ruiz, G. De Arcas

Research output: Contribution to conferencePaper

8 Citations (Scopus)

Abstract

The evolution in the past years of Machine learning techniques, as well as the technological evolution of computer architectures and operating systems, are enabling new approaches for complex problems in different areas of industry and research, where a classical approach is nonviable due to lack of knowledge of the problem's nature. A typical example of this situation is the prediction of plasma disruptions in Tokamak devices. This paper shows the implementation of a real time disruption predictor. The predictor is based on a support vector machine (SVM). The implementation was done under the MARTe framework on a six core x86 architecture. The system is connected in JET's Real time Data Network (RTDN). Online results show a high degree of successful predictions and a low rate of false alarms thus, confirming its usefulness in a disruption mitigation scheme. The implementation shows a low computational load, which in an immediate future will be exploited to increase the prediction's temporal resolution. © 2012 IEEE.
Original languageEnglish
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event2012 18th IEEE-NPSS Real Time Conference, RT 2012 - , United States
Duration: 1 Jan 2012 → …

Conference

Conference2012 18th IEEE-NPSS Real Time Conference, RT 2012
CountryUnited States
Period1/1/12 → …

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Information Systems

Cite this

Lopez, J. M., Vega, J., Alves, D., Dormido-Canto, S., Murari, A., Ramirez, J. M., ... De Arcas, G. (2012). Implementation of the disruption predictor APODIS in JET real time network using the MARTe framework. Paper presented at 2012 18th IEEE-NPSS Real Time Conference, RT 2012, United States. https://doi.org/10.1109/RTC.2012.6418168