Advanced Disruption Predictor Based On The Locked Mode Signal: Application To Jet

Vega, J.; Moreno, R.; Pereira, A.; Dormido-Canto, S.; Murari, A.

doi:10.22323/1.240.0028

Cited by 17 publications

(25 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although prediction performances of the proposed indicator are quite good, and also better than those reported in the recent literature [7], the analyses show the intrinsic limits of a possible disruption predictor based on the unique locked mode signal. Future works will be focused on using the proposed LM indicator together with other plasma features as inputs of a multi-signal disruption prediction system, optimised in terms of success rate, warning times and that be differentiated depending on different disruptions classes.…”

Section: Discussionmentioning

confidence: 60%

“…By analysing the JET data, it follows that, in the disruptive training set, all the discharges have a locked mode, but, for 1.85% of them, the locked mode signal starts to rise less than 10ms before the disruption. According to [7], in JET a detection is considered tardy if the alarm is triggered less than 10ms before the disruption time, where 10ms is minimum time need by JET mitigation systems to intervene. The analysis of the disruption test set showed that for 1.90% of disruptions there is no locked mode, and for as much as 19.52% of them the locked mode signal starts to rise less than 10ms from disruption time, hence too late to allow even a mitigation action.…”

Section: Analysis Of the Locked Mode Signalmentioning

confidence: 99%

“…It can be noted that the LM-ind obtains better results than BML/Bϑ(0,a), anticipating the correct predictions and, at the same time, limiting the premature detections. Indeed, according to [7], setting the limit between correct and premature alarms at 1.5s before the disruption time (cyan vertical dashed line in Figure 9), the achieved percentage of premature detections (PDs) are equal to 4.85% for LM-ind and 8.25% for BML/Bϑ(0,a) on the test set. Note that, from the plot of Figure 9, the percentage of the tardy detections can be also obtained depending on its definition (dashed magenta line).…”

Section: Machine-dependent Locked Mode Disruption Predictormentioning

confidence: 99%

“…These algorithms demonstrated to perform better than those that trigger an alarm only when the locked mode amplitude exceeds a prefixed threshold. Also in [7] time-frequency analysis was used to detect anomalies in the locked mode signal that trigger the disruption alarms on JET. Such predictor required a limited number of past experiments to set the alarm threshold.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A locked mode indicator for disruption prediction on JET and ASDEX upgrade

Sias

Cannas

Fanni

et al. 2019

Fusion Engineering and Design

View full text Add to dashboard Cite

The aim of this paper is to present a signal processing algorithm that, applied to the raw Locked Mode signal, allows us to obtain a disruption indicator in principle exploitable on different tokamaks. A common definition of such an indicator for different machines would facilitate the development of portable systems for disruption prediction, which is becoming of increasingly importance for the next tokamak generations. Moreover, the indicator allows us to overcome some intrinsic problems in the diagnostic system such as drift and offset. The behavior of the proposed indicator as disruption predictor, based on crossing optimized thresholds of the signal amplitude, has been analyzed using data of both JET and ASDEX Upgrade experiments. A thorough analysis of the disruption prediction performance shows how the indicator is able to recover some missed and tardy detections of the raw signal. Moreover, it intervenes and corrects premature or even wrong alarms due to, e.g., drifts and/or offsets.

show abstract

Section: Discussionmentioning

confidence: 60%

Section: Analysis Of the Locked Mode Signalmentioning

confidence: 99%

Section: Machine-dependent Locked Mode Disruption Predictormentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A locked mode indicator for disruption prediction on JET and ASDEX upgrade

Sias

Cannas

Fanni

et al. 2019

Fusion Engineering and Design

View full text Add to dashboard Cite

show abstract

“…Given the structural stability that they achieve by implementing the margins, SVMs are very powerful tools and have performed extremely well in the case of disruption prediction, as demonstrated in real time by APODIS [9,10]. Their hyperplane can therefore be considered a very good approximation of the boundary between the disruptive and not disruptive regions of the operational space.…”

Section:  mentioning

confidence: 99%

Adaptive predictors based on probabilistic SVM for real time disruption mitigation on JET

et al. 2018

View full text Add to dashboard Cite

Detecting disruptions with sufficient anticipation time is essential to undertake any form of remedial strategy, mitigation or avoidance. Traditional predictors based on machine learning techniques can be very performing if properly optimised but do not provide a natural estimate of the quality of their outputs. In this paper a new set of tools, based on probabilistic extensions of Support Vector Machines, are introduced and applied for the first time to JET data. The probabilistic output constitutes a natural qualification of the prediction quality and provides additional flexibility. Indeed the versatility of the developed techniques is such that different versions of the tools can be optimised to perform various tasks, from prediction for mitigation to avoidance or even classification. Large databases of disruptions, covering entire campaigns, are analysed, both for the case of the graphite and the ITER Like Wall. Success rates of the order of 97% with about 4.5% of false alarms can be easily achieved, satisfying even the requirements of the next generation of devices. The fact that the developed tools give the probability of disruption improves the interpretability of the results, provides an estimate of the predictor quality and gives new insights into the physics. Moreover, a probabilistic treatment permits to insert more easily these classifiers into general decision support and control systems.

show abstract