Dimensionnement d'un système d'aimants à fort champ pour tokamaks compacts

Abstract :

Magnets are at the core of tokamak design and performance. Initially built using copper, then low temperature superconductors (LTS), it is now foreseen to use advanced high temperature superconductors (HTS) to make tokamaks more compact at higher magnetic fields. Although the fusion power is significantly increased as implied by the scaling law , the larger fields generate intense electromagnetic forces on the magnet structures that reach the absolute material limits.

Considering these new mechanical challenges at the earliest stage of the design is crucial. Currently, a significant gap exists between analytical system codes (which lack detailed 3D structural fidelity, particularly regarding out-of-plane loads) and full-scale 3D finite element models (which carry prohibitive time and computational costs). This thesis bridges that gap by establishing the CIRCE toolbox, a comprehensive framework dedicated to the rapid and robust structural assessment of tokamak magnet systems. Utilizing a hybrid beam-shell approach, CIRCE accurately captures the global stiffness and complex load-bearing mechanism of toroidal field (TF) coils and their intercoil support structures. The code integrates dedicated geometry generation, regularized Biot-Savart magnetic field mapping, and reduced-order mechanical elements. This formulation allows for the rapid evaluation of both in-plane and out-of-plane constraints, enabling parametric explorations of the TF magnet structures with computational times reduced to seconds.

Beyond global structural architecture, the viability of a compact, high-field magnet design is inherently constrained by material limits. High-strength steels, such as CHSN01 (China Steel Number 01), offer the exceptional yield strength required to withstand massive static loads. However, the high strength can be associated with degraded fatigue performance due to the generation of high initial residual stresses and reduced fracture toughness. To refine the fatigue life assessment of the central solenoid (CS), this thesis couples the modeling of residual stresses with an original experimental campaign. X-ray diffraction and hole-drilling measurements of residual stresses in CHSN01 conductor jackets provide physically justified inputs for linear elastic fracture mechanics (LEFM) workflows, addressing critical bottlenecks in predicting the fatigue lifetime of the magnet.

Ultimately, the methodologies developed in this manuscript strengthen the early-stage design of magnetic confinement fusion devices. By providing fast structural assessments and empirical fatigue evaluations, this work ensures that complex mechanical constraints are integrated from the very beginning of the conceptual design phase.

Jury

• Mme Daniela BOSO - Professeure - Università di Padova - Rapporteure
• M. Alain RASSINEUX - Professeur - Université de Technologie de Compiègne - Rapporteur
• Mme Véronique AUBIN - Professeure - Centrale Supélec - Président
• M. Alfredo PORTONE - Senior researcher - Fusion for Energy - Examinateur
• M. Xabier SARASOLA - Senior researcher - École Polytechnique Fédérale de Lausanne - Examinateur
• M. Frédéric LEBON - Professeur - Aix-Marseille Université / LMA - Directeur de thèse
• M. Alexandre TORRE - Ingénieur de recherche - CEA Cadarache - Co-encadrant de thèse

.

La soutenance de thèse de Baptiste Boudes est prévue le 10 novembre à 10h00 au CEA-IRFM, salle René Gravier, Bâtiment 506