Stochastic modeling of turbulent plasma fluctuations applied to the TCABR tokamak

• Ano de defesa: 2022
• Autor(a) principal: Zurita, Martim
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: corrente de saturação iônica; electrostatic turbulence; física de plasmas; ion saturation current; modelagem estocástica; plasma physics; stochastic modeling; tokamaks; turbulência eletrostática
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/43/43134/tde-04042022-130632/

In toroidal devices known as tokamaks, high-temperature plasmas are confined by intense magnetic fields. Nevertheless, this confinement is deteriorated by turbulence at the edge of the devices. This turbulence has an intermittent behavior with the presence of high-amplitude bursts. To describe local measurements of density with bursts, a stochastic pulse train model (SPTM) has been developed since the last decade. For such a model, different categories of background signals have been considered in the literature, namely, backgrounds with Gaussian noises (correlated and uncorrelated) or with small-amplitude pulses. However, until now these models with different background signals weren\'t simultaneously compared to an experiment. Moreover, there isn\'t a fitting method for the SPTM that can evaluate all its parameters in a unified and objective way. The present dissertation aims to fulfill these two gaps. Having created the SPTM fit, we applied it to the TCABR tokamak. For this analysis, we utilized measurements of ion saturation current, a signal proportional to the local plasma density. In addition, we introduced to the context of the SPTM two non-linear tools: the complexity-entropy diagram and the determinism from recurrence quantification analysis. With them and the frequency spectrum, we concluded that, for the analyzed experiment, the model with a pulse background described the structure of plasma density fluctuations better than the models with Gaussian noise.