Simulacija magnetnega vodenja oblike plazme v tokamaku ITER

Knap, Matic (author), Širca, Simon (mentor), Gerkšič, Samo (comentor)

Magistrsko delo obravnava magnetno vodenje oblike plazme v tokamaku ITER. Najprej zapišemo osnovni magnetni model plazme, ki se uporablja pri simulaciji za vodenje. V kasnejših poglavjih se posvetimo samemu algoritmu za vodenje toka in oblike plazme. Predstavimo implementacijo magnetnega modela plazme in prediktivno metodo vodenja. Osredotočimo se na prilagoditve, ki so bile izvedene na obstoječem algoritmu za reševanje optimizacijskih problemov. Reševanje regulacijskega problema prediktivnega vodenja metodo poteka v dveh delih. Najprej model pripravimo za regulator z zaporedjem operacij, s katerim dobimo dosti bolj kompakten model, ki še vedno dobro opisuje dinamiko sistema. S tem zagotovimo robustnost in točnost metode. Prediktivno metodo pretvorimo v optimizacijski problem kvadratnega programiranja, ki ga rešujemo s hitro gradientno metodo prvega reda. Delo zajema tudi optimizacijo izvajanja programa. Sledi prikaz rezultatov simulacij, kjer prikažemo, da je izbrana metoda s prilagoditvami časovno ugodnejša od trenutnih komericalnih metod ob enaki natančnosti.

• Language: Slovenian
• Keywords: tokamak, plazma, regulacijski sistem, prediktivni regulator, ITER
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FMF - Faculty of Mathematics and Physics
• Year: 2017
• PID: 20.500.12556/RUL-97780
• COBISS.SI-ID: 3148132
• Publication date in RUL: 10.11.2017

Secondary language

• Language: English
• Title: Simulation of the Magnetic Plasma Shape Control for the tokamak ITER

Abstract

The main subject of this master thesis is the current and shape control of plasma in the ITER tokamak. Firstly, we derive the basic magnetic model of a plasma, which is used for simulations. Later on we describe an algorithm for plasma current and shape control. The model predictive control (MPC) method is presented and the actual implementation of the magnetic model of plasma. We focus on modifications to the algorithm. Solving the plasma shape control problem with model predictive control is divided into two parts. Firstly, we transform the model by applying algebra to the problem. This ensures us robustness and accuracy of the algorithm. The MPC problem is converted to the optimization problem of quadratic programming (QP) which is solved using the fast gradient method. The thesis includes also the optimization of the execution time. Later on we present results of the simulations and show that the chosen method with modifications is more time efficient than currently available methods.

• Keywords: tokamak, plasma, systems control, model predictive control, ITER