Sias Giuliana
UniCa Ateneo Docenti e ricercatori Giuliana Sias Curriculum

Giuliana Sias

Dipartimento di Ingegneria elettrica ed elettronica

Carriera accademica

  • Da dicembre 2020, Professoressa Associata di Elettrotecnica (SSD ING-IND/31) presso il Dip. di Ing. Elettrica ed Elettronica, Università degli Studi di Cagliari.
  • Dicembre 2010 – dicembre 2020, Ricercatrice a tempo indeterminato, settore Elettrotecnica (SSD ING-IND/31) presso il Dip. di Ing. Elettrica ed Elettronica, Università degli Studi di Cagliari.
  • Marzo 2007 - dicembre 2010, Postdoc, settore Elettrotecnica (SSD ING-IND/31) presso il Dip. di Ing. Elettrica ed Elettronica, Università degli Studi di Cagliari

Formazione

  • Marzo 2007, Università degli Studi di Padova. Dottorato di ricerca in Ing. Elettrotecnica, con tesi dal titolo: A disruption prediction system for ASDEX Upgrade based on Neural Networks, con certificazione congiunta di “Doctor Europaeus” allegata al titolo nazionale
  • Luglio 2003, Università degli Studi di Cagliari. Laurea in Ingegneria Elettrica con votazione di 110/110 e lode, con tesi dal titolo: Uno strumento virtuale per lo studio dei disturbi di Power Quality mediante Trasformata Wavelet

Attività didattica

Titolare dei seguenti corsi:

  • Elettromagnetismo Applicato all'Ingegneria Elettrica ed Energetica. Università degli Studi di Cagliari, Lauree magistrale in Ingegneria Elettrica (9CFU) e in Ingegneria Energetica (6CFU). A.A. da 2018-2019 ad oggi.
  • Elettrotecnica (5 CF), Università degli Studi di Cagliari, Laurea in Ingegneria Civile. A.A. dal 2010-2011 al 2018-2019.
  • Laboratorio di Elettrotecnica (5 CFU). Università degli Studi di Cagliari, Laurea in Ingegneria per l'Ambiente e il Territorio. A.A. dal 2011-2012 al 2014-2015.

Temi di ricerca

Applicazione di algoritmi di intelligenza artificiale e tecniche di data mining, pre e post-processing per la classificazione, la predizione, l’ottimizzazione e la diagnostica nel campo della fusione termonucleare controllata, gestione delle smart grid e applicazioni mediche.

a) Algoritmi di Machine Learning per la predizione e l'avoidance delle disruzioni macchine sperimentali a fusione termonucleare controllata ASDEX Upgrade e JET.  Applicazione e l’implementazione di algoritmi basati sui dati in grado di predire tempestivamente eventi endemici nelle macchine a fusione termonucleare controllata di tipo tokamak, chiamate disruzioni, che causano una repentina perdita di confinamento del plasma. In questo contesto, sono stati applicati innovativi algoritmi di Machine Learning in grado di rilevare in maniera tempestiva gli eventi chiave delle catene di fenomeni che portano a disruzione, così da poter intraprendere appropriate azioni di controllo per ricondurre il plasma in una situazione stabile (vedi progetti 6, 7, 10, 11, 12, 15, 16, 19, 21). Inoltre, sono stati sviluppati predittori neurali in grado di attivare in maniera opportuna i sistemi di protezione dai danni alla struttura della macchina che seguito di una disruzione, possono essere causati dall’elevato e brusco rilascio di energia (vedi progetti da 23 a 29). Le attività di ricerca sono sviluppate in collaborazione con i ricercatori dell'IPP Max-Planck-Institut für Plasmaphysik di Garching b. München (DE) e il Culham Center for Fusion Energy (CCFE), presso i quali si trovano rispettivamente i tokamak ASDEX Upgrade e JET.

b) Reti neurali artificiali per la diagnostica delle immagini nella fusione termonucleare controllata. Applicazione di reti neurali tradizionali e di algoritmi di deep learning per l'analisi e la caratterizzazione di eventi termici (strike-lines) nei moduli del divertore dello stellarator W7-X (vedi progetti 8, 9, 13, 14, 17, 18, 20); per la ricostruzione inversa del flusso termico sul fronte delle piastre del calorimetro strumentale del prototipo della sorgente di ioni negativi (SPIDER) dell’iniettore di neutri del futuro reattore sperimentale ITER, a partire dalle misure di temperature sul retro, effettuate camere IR; per il rilevamento dei filamenti da immagini 2D di telecamere veloci nel visibile, per il tokamak sferico MAST Upgrade. Le attività di ricerca sono sviluppate in collaborazione con i ricercatori dell'IPP Max-Planck-Institut für Plasmaphysik, Greifswald (DE), dove è situato W7-X, dell’Istituto Gas Ionizzati di Padova, dove di trova l’esperimento SPIDER, e il Culham Center for Fusion Energy (CCFE) dove è situato MAST Upgrade.

c) Tecniche di data mining per la creazione di un database di eventi transienti rilevanti per il di futuro reattore a fusione EU-DEMO. Applicazione di tecniche di data mining, di pre e post-elaborazione per la creazione di un database multi-macchina di perturbazioni che portano all’instabilità verticale della colonna di plasma, in esperimenti rilevanti per il rettore DEMO di prossima costruzione in Europa (vedi progetti 1, 3, 4, 5). Le attività di ricerca sono sviluppate in collaborazione con i ricercatori dell’EUROfusion-Programme Management Unit, Garching (DE).

d) Caratterizzazione del rumore bianco e modellazione delle sollecitazioni termomeccaniche delle bobine Pick-up per il tokamak DTT. Valutazione degli errori sistematici sulla ricostruzione inversa della posizione del plasma dovuti all'effetto del rumore bianco sulle misure di campo, effettuate tramite bobine pick-up in-vessel. Analisi dello stress termomeccanico delle bobine pick-up ex-vessel in torlon/rame tramite codice agli elementi finiti (vedi progetto 2).

e) Intelligenza artificiale applicata alla gestione delle smart-grid. Predizione innovativa dei consumi elettrici e sul Non-Intrusive Load Monitoring (NILM) di dispositivi elettrici, integrati in una piattaforma IoT, mediante modelli basati sui dati.

f) Deep learning per applicazioni mediche. Applicazione di algoritmi di deep-learning per la rilevazione automatica delle crisi nell'epilessia frontale notturna.

Principali progetti di ricerca

  1. Gennaio - Dicembre 2022 - EUROfusion program, Work package “DES-FS.PLA.S-T014-D002”. Deliverable owner per ENEA-Univ. di Cagliari nella task PLA.S.02-06.2, titolo: Disruption Expert 2022.
  2. Gennaio - Dicembre 2021 – DTT (Divertor Tokamak Test) 2021 Task “DMA_Diagnostics_Magnetic”. Sub task owner per Univ. della Tuscia - Univ. di Cagliari, work package ID 4.10.4.1.11_001 White noise characterization and FEM modelling of thermo-mechanical stresses of Pick-up coils.
  3. Gennaio - Dicembre 2021 - EUROfusion program, Work package “DES-FS.PLA.S-T014-D001". Deliverable owner per ENEA-Univ. di Cagliari per il task PLA.S.02-06.2, titolo: Disruption Expert 2021.
  4. Gennaio - Luglio 2020 - EUROfusion program, Work package "PMI5.3 Demo Physics design integration". Deliverable owner per ENEA-Univ. di Cagliari per il task PMI-5.3.2-T028, titolo: Plasma Perturbation database in DEMO relevant scenarios.
  5. Gennaio - Dicembre 2019 - EUROfusion program, Work package “PMI5.3 Demo Physics design integration” di EUROfusion. Deliverable owner per ENEA-Univ. di Cagliari, task 5.3.2-T022, titolo: “JET and ASDEX Upgrade plasma perturbation database in DEMO relevant scenarios”.
  6. 2021 EUROfusion program on Tokamak exploitation, WPTE-2021. Topic RT04: Disruption avoidance and control for ITER and DEMO
  7. 2021 EUROfusion program on JET experimental campaigns, WPJET1-2021. Task M21-03: Baseline scenario for high fusion performance in DT, Task T17-03: MHD analysis in support of the scenario development
  8. 2020-21- EUROfusion program on W7-X OP1.2b. Image processing - pattern recognition OP1, (2020) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.  
  9. 2020 -21 - EUROfusion program on W7-X OP1.2b. Analysis of strike-line control in OP1, (2020) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.  
  10. 2020 - EUROfusion program on JET experimental campaigns, WPJET1-2020. Task M18-04: Plasma termination and disruption avoidance for scenarios
  11. 2020 - EUROfusion program on medium-size tokamaks, WPMST1-2020. Topic T06-AUG: Disruption Prediction and Avoidance.
  12. 2019 - EUROfusion program on JET experimental campaigns, WPJET1-2019. Topic 06-AUG: Disruption Prediction and Avoidance
  13. 2019 - EUROfusion program on W7-X OP1.2b. Image processing - pattern recognition OP1, (2019) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1,
  14. 2019 - EUROfusion program on W7-X OP1.2b. Analysis of strike-line control in OP1, (2019) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.
  15. 2018-2019 - EUROfusion program on JET experimental campaigns, WPJET1-2018. Task M18-04: Plasma termination and disruption avoidance for scenarios
  16. 2018 - EUROfusion program on Medium-Size Tokamaks, WPMST1-2020. Topic T06-AUG: Disruption Prediction and Avoidance.
  17. 2018 - EUROfusion program for W7-X EXPERIMENTAL CAMPAIGN OP1.2(b). Task S1.P2.T7: Image processing – pattern recognition
  18. 2018 - EUROfusion program for W7-X EXPERIMENTAL CAMPAIGN OP1.2(b). Task: Imaging Software WP18.S1. B3.T4 2018, Spatial camera calibration and strike-line characterization
  19. 2017 - EUROfusion program on Medium Size Tokamaks, WPMST1 (2017). Topic T7-AUG: Disruption prediction, routine avoidance and mitigation in a broad set of scenarios
  20. 2017 - EUROfusion program on W7-X EXPERIMENTAL CAMPAIGN OP1.2(a). Experimental Campaign, Modification of strike line with control coils
  21. 2017 - EUROfusion program on JET experimental campaigns, WPJET1-2017. Task 17-14: Disruption avoidance and plasma termination
  22. 2015-2016 - EUROfusion program on JET experimental campaigns, WPJET1 2015-216. Task T15-3: Disruption prevention and avoidance schemes for JET
  23. 2015 - EUROfusion Work Program on Medium Size tokamaks, WPMST1 2015-16. Task T15-03: Disruption prediction
  24. 2014 - EUROfusion Work Program on JET experimental campaigns, analysis of JET C31-C34 experiments and tasks -2014. Task T13-23
  25. 2014 - EUROfusion Work Program on Medium Size Tokamaks, WPMST1-2014. Task AUG14-1.1-1
  26. 2013 - EFDA 2013 Work Program Topical Area A7, Disruptions Prediction Avoidance Mitigation and Consequences, founded by EURATOM
  27. 2012 - EFDA 2012 Work Program Topical Area A7 "Disruptions Prediction Avoidance Mitigation and Consequences, founded by EURATOM
  28. Progetto PRIN 2008. Modelli predittivi per la mitigazione di disruzioni in macchine ITER-like per la fusione termonucleare controllata.
  29. 2007-2013 - Development of disruption protection tools for Tokamak reactors, founded by the EURATOM and ENEA agreement.

Premi e riconoscimenti

  • Highlight dell'anno 2015 della rivista internazionale Plasma Physics and Controlled Fusion per l'articolo: Automatic disruption classification in JET with the ITER-Like Wall, B Cannas, P.C. de Vries, A Fanni, A Murari, A Pau, G Sias, and JET EFDA Contributors, Plasma Physics and Controlled Fusion, International Atomic Energy Agency (IAEA) Ed., 16.10.2015, 57 125003
  • Premio di €10000 attribuito dal Sistema premiale della ricerca, di cui all’art. 13 della L.R. n. 7/2007 per progetti di rilievo internazionale, titolo “MST1 Medium-Size Tokamak Campaigns - Work Program 2015-16” - Contributo Premiale Regione Autonoma della Sardegna L.R. 7/2007, Anno 2015
  • Premio come pubblicazione rilevante nell’area scientifica “Tecnologie applicate alle macchine per o studio dei plasmi” all’articolo: Delogu Rita S., Montisci A., Pimazzoni A., Serianni G., Sias G. (2019). Neural Network based prediction of heat flux profiles on STRIKE. Fusion Engineering and Design 146, 2307-2313. Assegnato dalla commissione per l’attribuzione del premio risultato 2019 per il personale con qualifica di ricercatore e tecnologo del Consorzio RFX.

Academic career

  • Currently, Associate Professor of circuit theory (SSD ING-IND/31) at Electrical and Electronic Engineering - Dep. University of Cagliari, Italy.
  • December 2010- December 2020, Research Scientist (permanent position) of circuit theory (SSD ING-IND/31) at Electrical and Electronic Engineering Dep. - University of Cagliari, Italy
  • March 2007 - December 2010, Postdoc Researcher of circuit theory (SSD ING-IND/31) at Electrical and Electronic Engineering Dep. - University of Cagliari, Italy

Education

  • June 2003: “Laurea” degree (magna cum laude) in Electrical Engineering, University of Cagliari, Italy, title of Master Thesis: A virtual tool for studying Power Quality disturbances using Wavelet Transform. 
  • March 2007: PhD in Electrical Engineering with “Doctor Europaeus” certificate, University of Padova, Italy. PhD Thesis: A disruption prediction system for ASDEX Upgrade based on Neural Networks.

Teaching Activity

In charge of the following courses:

  • Applied Electromagnetism. University of Cagliari, Cagliari (IT), Master’s degrees in Electrical Engineering (9 CFU) and Energetic Engineering (6 CFU). Academic years, from 2018-2019 to currently.
  • Circuits Theory (5 CF), University of Cagliari, Cagliari (IT), Bachelor’s degree in Civil Engineering. Academic years, from 2010-2011 to 2018-2019.
  • Circuits Theory Laboratory (5 CFU). University of Cagliari, Cagliari (IT), Bachelor’s degree in Environmental Engineering. Academic years, from 2011-2012 to 2014-2015.

Research Topics

The research activity is focused on the application of artificial intelligence algorithms and the development of data mining, pre and post-processing techniques for classification, prediction, optimization and diagnostics in the controlled fusion, smart grid management and medicine fields.

a) Machine Learning algorithms for disruption prediction and avoidance at ASDEX Upgrade and JET Tokamaks. Application and implementation of data-driven algorithms able to promptly detect disruptions to be handle in real-time with appropriate actions. In this context, Artificial Neural Networks predictors have been developed to activate the disruption protection systems (see project from 23 to 28) to prevent the machine from damages due to the abrupt loss of plasma energy confinement.  Moreover, innovative Machine Learning algorithms have been applied for the detection of key events in disruptive paths to activate the discharge control system to seek to recover the plasma to the previous stable state (see project 6, 7, 10, 11, 12, 15, 16, 19, 21). The research activities have been developed in collaboration with the researchers from the ASDEX Upgrade team and the JET contributors. Research activities are developed in collaboration with researchers from Culham Centre for Fusion Energy (CCFE), Abingdon (UK), and at IPP Max-Planck-Institut für Plasmaphysik, Garching b. München (DE), where are located the controlled fusion experimental machines JET and ASDEX Upgrade, respectively.

b) Artificial Neural Network applications for image diagnostics in controlled fusion. Application of Multi-layer perceptron and Deep Neural Networks for the analysis and the characterization of thermal events (strike-lines) in the W7-X stellarator divertor modules (see projects 8, 9, 13, 14, 17, 18, 20); inverse reconstruction of the thermal flux from IR temperature measurements on the instrumental calorimeter tiles for the ITER negative ion beam source (SPIDER); filament detections from 2D fast camera images for the spherical tokamak MAST Upgrade. Research activities are developed in collaboration with researchers from IPP Max-Planck-Institut für Plasmaphysik, Greifswald (DE), where is located W7-X, the Istituto Gas Ionizzati, Padova (IT), where is located the experiment SPIDER, and the Culham Center for Fusion Energy (CCFE), where is located MAST Upgrade.

c) Data-mining techniques for the creation of a DEMO relevant off-normal event database. Application of data mining, pre and post-processing techniques for the creation of a multi-machine database of plasma perturbations inducing vertical displacement events (VDEs), in relevant DEMO experiments (see projects 1, 3, 4, 5). Research activities are developed in collaboration with researchers from EUROfusion-Programme Management Unit, Garching (DE).

d) White noise characterization and modelling of thermo-mechanical stresses of Pick-up coils. Assessment of the systematic error on the inverse plasma position and shape reconstruction due to the white noise effect on the in-vessel pick-up coil measurements. Thermo-mechanical stress analysis of torlon/copper ex-vessel pick-up coils using finite element code (see project 2).

e) Artificial intelligence applied to smart-grid management. Innovative prediction of electricity consumption and the Non-Intrusive Load Monitoring (NILM) of electrical devices, integrated into a IoT platform, by data-driven models.

f) Deep learning for medical applications. Application of deep learning algorithms for the automatic detection of epileptic seizures in nocturnal frontal epilepsy.

Main research projects

  1. 1° January -31 December 2022 - EUROfusion program, Work package “DES-FS.PLA.S-T014-D002”. Deliverable owner for ENEA-Univ. of Cagliari in the task PLA.S.02-06.2, title: Disruption Expert 2022.
  2. 1° January -31 December 2021 – DTT (Divertor Tokamak Test) 2021 Task “DMA_Diagnostics_Magnetic”. Sub task owner for La Tuscia - Univ. di Cagliari, work package ID 4.10.4.1.11_001 White noise characterization and FEM modelling of thermo-mechanical stresses of Pick-up coils.
  3. 1° January -31 December 2021 - EUROfusion program, Work package “DES-FS.PLA.S-T014-D001". Deliverable owner for ENEA-Univ. di Cagliari in the task PLA.S.02-06.2, title: Disruption Expert 2021.
  4. 1° January -31 July 2020 - EUROfusion program, Work package "PMI5.3 Demo Physics design integration". Deliverable owner for ENEA-Univ. di Cagliari in the task PMI-5.3.2-T028, title: Plasma Perturbation database in DEMO relevant scenarios.
  5. 1° January -31 December 2019 - EUROfusion program, Work package “PMI5.3 Demo Physics design integration” di EUROfusion. Deliverable owner for ENEA-Univ. di Cagliari, task 5.3.2-T022: “JET and ASDEX Upgrade plasma perturbation database in DEMO relevant scenarios”.
  6. Member of research team:
  7. 2021 EUROfusion program on Tokamak exploitation, WPTE-2021. Topic RT04: Disruption avoidance and control for ITER and DEMO
  8. 2021 EUROfusion program on JET experimental campaigns, WPJET1-2021. Task M21-03: Baseline scenario for high fusion performance in DT, Task T17-03: MHD analysis in support of the scenario development
  9. 2020-21- EUROfusion program on W7-X OP1.2b. Image processing - pattern recognition OP1, (2020) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.  
  10. 2020 -21 - EUROfusion program on W7-X OP1.2b. Analysis of strike-line control in OP1, (2020) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.  
  11. 2020 - EUROfusion program on JET experimental campaigns, WPJET1-2020. Task M18-04: Plasma termination and disruption avoidance for scenarios
  12. 2020 - EUROfusion program on medium-size tokamaks, WPMST1-2020. Topic T06-AUG: Disruption Prediction and Avoidance.
  13. 2019 - EUROfusion program on JET experimental campaigns, WPJET1-2019. Topic 06-AUG: Disruption Prediction and Avoidance
  14. 2019 - EUROfusion program on W7-X OP1.2b. Image processing - pattern recognition OP1, (2019) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1,
  15. 2019 - EUROfusion program on W7-X OP1.2b. Analysis of strike-line control in OP1, (2019) S1 Work package, 19-20 Preparation and Exploitation of W7-X Campaigns. Task S1.X2.A.T2: Analysis of strike-line control in OP1.
  16. 2018-2019 - EUROfusion program on JET experimental campaigns, WPJET1-2018. Task M18-04: Plasma termination and disruption avoidance for scenarios
  17. 2018 - EUROfusion program on Medium-Size Tokamaks, WPMST1-2020. Topic T06-AUG: Disruption Prediction and Avoidance.
  18. 2018 - EUROfusion program for W7-X EXPERIMENTAL CAMPAIGN OP1.2(b). Task S1.P2.T7: Image processing – pattern recognition
  19. 2018 - EUROfusion program for W7-X EXPERIMENTAL CAMPAIGN OP1.2(b). Task: Imaging Software WP18.S1. B3.T4 2018, Spatial camera calibration and strike-line characterization
  20. 2017 - EUROfusion program on Medium Size Tokamaks, WPMST1 (2017). Topic T7-AUG: Disruption prediction, routine avoidance and mitigation in a broad set of scenarios
  21. 2017 - EUROfusion program on W7-X EXPERIMENTAL CAMPAIGN OP1.2(a). Experimental Campaign, Modification of strike line with control coils
  22. 2017 - EUROfusion program on JET experimental campaigns, WPJET1-2017. Task 17-14: Disruption avoidance and plasma termination
  23. 2015-2016 - EUROfusion program on JET experimental campaigns, WPJET1 2015-216. Task T15-3: Disruption prevention and avoidance schemes for JET
  24. 2015 - EUROfusion Work Program on Medium Size tokamaks, WPMST1 2015-16. Task T15-03: Disruption prediction
  25. 2014 - EUROfusion Work Program on JET experimental campaigns, analysis of JET C31-C34 experiments and tasks -2014. Task T13-23
  26. 2014 - EUROfusion Work Program on Medium Size Tokamaks, WPMST1-2014. Task AUG14-1.1-1
  27. 2013 - EFDA 2013 Work Program Topical Area A7, Disruptions Prediction Avoidance Mitigation and Consequences, founded by EURATOM
  28. 2012 - EFDA 2012 Work Program Topical Area A7 "Disruptions Prediction Avoidance Mitigation and Consequences, founded by EURATOM
  29. PRIN 2008. Predictive models for disruption mitigation in in ITER-like machines for controlled thermonuclear fusion.
  30. 2007-2013 - Development of disruption protection tools for Tokamak reactors, founded by the EURATOM and ENEA agreement.

Awards

  • Highlight of the year 2015 for the article: Automatic disruption classification in JET with the ITER-Like Wall, 10.1088/0741-3335/57/12/125003. Plasma Physics and Controlled Fusion, International Atomic Energy Agency (IAEA)
  • €10000 from the Sardinia Research reward system, year 2015, art. 13 L.R. n. 7/2007, for the international -project: “MST1 Medium-Size Tokamak Campaigns - Work Program 2015-16”.
  • Awards as relevant publication for the article: Neural network based prediction of heat flux profiles on STRIKE, 10.1016/j.fusengdes.2019.03.178.  Awarded by the Consorzio RFX commission for the results of the year 2019 in the field: Technologies applied to machines for the plasma study.

Questionario e social

Condividi su:
Impostazioni cookie