Nuclear fusion represents one of the most promising sources of energy by virtue of its high energy density, large abundance of the fuels and its intrinsic safety. The process in- volves two light nuclei that fuse, producing energy that can be gathered and converted in electricity. The temperature necessary to achieve nuclear fusion forces the matter to the state of plasma, a system of charged particles, globally neutral, with collective behaviour. The tokamaks are machines used to confine a plasma of hydrogen isotopes, by means of a combination of magnetic fields, allowing the plasma to be heated until the energy of the particles overcomes the coulomb barrier and the nuclei fuse. The plasma exhibits instabilities that develop over various time and spatial scales. This thesis focuses on a particular kind of resistive magnetohydrodynamical instability, called the neoclassical tearing mode ((N)TM). The onset of the (N)TM manifests with a change in the topology of the confining magnetic field, forming a magnetic island, which can grow, reducing the performances and possibly leading to a complete loss of confinement of the plasma (disruption). The investigation of the (N)TM involves different aspects: the trigger of the mode, its evolution and control, and the role (N)TM can play in disrup- tive phenomena. In this work different aspects, addressed above, have been investigated exploiting both experimental data from the Joint European Torus (JET) and predictive sim- ulations of (N)TM evolution and control based on available scenarios from the Divertor Test Tokamak (DTT). After a general introduction (Chapter 1), and a description of DTT and JET (Chapter 2), a detailed theoretical analysis related to the physics of (N)TM is set in Chapter 3. The topic of control of (N)TM is covered in Chapter 4. The (N)TM evolution can be controlled injecting powerful Electron Cyclotron Waves that drives a current within the is- land. Predictive simulations in DTT full power scenario were carried out to address some requirements for the design of the electron cyclotron systems and investigating possible suppression strategies. The onset of the pre-disruptive (N)TM has been modelled in JET disruptive plasma. A physical picture where the ion polarization current covers a role in the trigger of tear- ing instabilities has been proposed in a published work (Appendix A), and introduced in Chapter 5. The topic of the early identification of disruptive conditions with deep learning is ad- dressed in Chapter 6 using a dataset composed by JET disruptive plasma preceded by the onset of an (N)TM. The classification of a deep neural network has been analyzed via explainable artificial intelligence (XAI) methods, to verify if the network can distinguish between different disruptive patterns that involve the trigger of a 2/1 mode. Section 6.4 presents the published work (reported in Appendix B)
La fusione nucleare rappresenta una delle fonti di energia più promettenti grazie alla sua alta densità energetica, alla grande abbondanza dei combustibili e alla sua intrinseca sicurezza. Il processo prevede la fusione di due nuclei leggeri, producendo energia che può essere raccolta e convertita in elettricità. La temperatura necessaria per ottenere la fusione nucleare porta la materia allo stato di plasma, un sistema di particelle cariche, globalmente neutro, con un comportamento collettivo. I tokamak sono macchine utilizzate per confinare un plasma di isotopi di idrogeno, attraverso una combinazione di campi magnetici, permettendo al plasma di riscaldarsi fino a che l’energia delle particelle supera la barriera coulombiana e i nuclei si fondono. Il plasma manifesta instabilità che si sviluppano su varie scale temporali e spaziali. Questa tesi si concentra su un particolare tipo di instabilità magnetoidrodinamica resistiva, chiamata modo tearing neoclassico ((N)TM). L’insorgenza del (N)TM si manifesta con un cambiamento nella topologia del campo magnetico confinante, formando un’isola magnetica, che può crescere, riducendo le prestazioni e portando a una completa perdita del confinamento del plasma (disruzione). L’indagine del (N)TM coinvolge diversi aspetti: il trigger del modo, la sua evoluzione e controllo, e il ruolo che il (N)TM può avere nei fenomeni di disruption. In questo lavoro sono stati indagati diversi aspetti, sopra citati, sfruttando sia i dati sperimentali del Joint European Torus (JET) sia le simulazioni predittive dell’evoluzione e controllo del (N)TM basate sugli scenari disponibili del Divertor Test Tokamak (DTT). Dopo un’introduzione generale (Capitolo 1) e una descrizione di DTT e JET (Capitolo 2), nel Capitolo 3 viene presentata un’analisi teorica dettagliata relativa alla fisica del (N)TM. Il tema del controllo del (N)TM è trattato nel Capitolo 4. L’evoluzione del (N)TM può essere controllata iniettando electron cyclotron waves che generano una corrente all’interno dell’isola. Sono state condotte simulazioni predittive nello scenario a piena potenza di DTT per definire alcuni requisiti per il design dei sistemi a ciclotrone di elettroni e indagare possibili strategie di soppressione dell’isola. L’insorgenza del (N)TM pre-disruptivo è stata modellata in plasma disruttivi di JET. È stata proposta una spiegazione fisica in cui la corrente di polarizzazione ionica gioca un ruolo nel trigger delle instabilità tearing in un lavoro pubblicato (Appendice A) e introdotta nel Capitolo 5. Il tema dell’identificazione precoce delle condizioni disruptive tramite deep learning è trattato nel Capitolo 6, utilizzando un dataset composto da plasmi di JET disruttivi, preceduti dall’insorgenza di un (N)TM. La classificazione di una deep neural network è stata analizzata tramite metodi di explainable artificial intelligence (XAI), per verificare se la rete riesce a distinguere tra diversi pattern disruttivi che coinvolgono l’innesco di un mode 2/1. La Sezione 6.4 presenta il lavoro pubblicato (riportato in Appendice B).
(2025). Investigation of the (neoclassical) tearing instabilities from the trigger to the disruption in JET and DTT. (Tesi di dottorato, , 2025).
Investigation of the (neoclassical) tearing instabilities from the trigger to the disruption in JET and DTT
BONALUMI, LUCA
2025
Abstract
Nuclear fusion represents one of the most promising sources of energy by virtue of its high energy density, large abundance of the fuels and its intrinsic safety. The process in- volves two light nuclei that fuse, producing energy that can be gathered and converted in electricity. The temperature necessary to achieve nuclear fusion forces the matter to the state of plasma, a system of charged particles, globally neutral, with collective behaviour. The tokamaks are machines used to confine a plasma of hydrogen isotopes, by means of a combination of magnetic fields, allowing the plasma to be heated until the energy of the particles overcomes the coulomb barrier and the nuclei fuse. The plasma exhibits instabilities that develop over various time and spatial scales. This thesis focuses on a particular kind of resistive magnetohydrodynamical instability, called the neoclassical tearing mode ((N)TM). The onset of the (N)TM manifests with a change in the topology of the confining magnetic field, forming a magnetic island, which can grow, reducing the performances and possibly leading to a complete loss of confinement of the plasma (disruption). The investigation of the (N)TM involves different aspects: the trigger of the mode, its evolution and control, and the role (N)TM can play in disrup- tive phenomena. In this work different aspects, addressed above, have been investigated exploiting both experimental data from the Joint European Torus (JET) and predictive sim- ulations of (N)TM evolution and control based on available scenarios from the Divertor Test Tokamak (DTT). After a general introduction (Chapter 1), and a description of DTT and JET (Chapter 2), a detailed theoretical analysis related to the physics of (N)TM is set in Chapter 3. The topic of control of (N)TM is covered in Chapter 4. The (N)TM evolution can be controlled injecting powerful Electron Cyclotron Waves that drives a current within the is- land. Predictive simulations in DTT full power scenario were carried out to address some requirements for the design of the electron cyclotron systems and investigating possible suppression strategies. The onset of the pre-disruptive (N)TM has been modelled in JET disruptive plasma. A physical picture where the ion polarization current covers a role in the trigger of tear- ing instabilities has been proposed in a published work (Appendix A), and introduced in Chapter 5. The topic of the early identification of disruptive conditions with deep learning is ad- dressed in Chapter 6 using a dataset composed by JET disruptive plasma preceded by the onset of an (N)TM. The classification of a deep neural network has been analyzed via explainable artificial intelligence (XAI) methods, to verify if the network can distinguish between different disruptive patterns that involve the trigger of a 2/1 mode. Section 6.4 presents the published work (reported in Appendix B)File | Dimensione | Formato | |
---|---|---|---|
phd_unimib_777689.pdf
accesso aperto
Descrizione: Investigation of the (neoclassical) tearing instabilities from the trigger to the disruption in JET and DTT
Tipologia di allegato:
Doctoral thesis
Dimensione
10.98 MB
Formato
Adobe PDF
|
10.98 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.