As the final step of my second year of a master’s program in magnetic fusion, I had the great opportunity to conduct my six-month master’s thesis at the Swiss Plasma Center (SPC) in Switzerland. Welcomed within one of the laboratories at EPFL, I worked on the TCV (Tokamak à Configuration Variable), a unique tokamak capable of shaping the plasma into various configurations within its vacuum vessel.
My research focused on the startup phase of the tokamak discharge, a crucial period lasting ~ 100 milliseconds in TCV experiments. During this phase, the plasma is created and heated sufficiently to sustain a plasma current. However, under certain conditions, runaway electrons (highly energetic) can impact the current ramp-up and potentially damage the vessel. Despite its importance, the startup phase relies heavily on operator experience rather than numerical modeling, which will be essential for optimizing startup procedures in large-scale fusion experiments such as ITER.
To improve our understanding of this phase and mitigate its challenges, one of the few existing simulation tools, STREAM (STartup Runaway Electron Analysis Model), has been developed. STREAM is a 0D model that estimates key fusion parameters, including the energy confinement time, using the Bohm diffusion coefficient. However, previous studies have shown that the Bohm-based estimate of τE\tau_EτE deviates significantly from experimental measurements, with discrepancies reaching up to a factor of 10³.
The primary goal of my research was to compute the energy confinement time during the startup phase using experimental data from TCV and compare these results with the predictions of the Bohm-based model used in STREAM. The analysis focused specifically on ohmically heated plasmas without the influence of runaway electrons. Additionally, my work initially aimed to refine the STREAM model by evaluating its predictions against other well-established scaling laws.
However, as the project evolved, unexpected challenges arose, like determining the most suitable approach for computing plasma conductivity during the start up phase—either the neoclassical or classical one. With the guidance of my supervisor, we observed that during the startup phase, the neoclassical approach appeared to be more relevant. While this shift in focus meant that some initial objectives could not be fully achieved within the internship period, the work laid a strong foundation for further studies and opened new directions for improving startup modeling.
This internship was one of the most fulfilling experiences I have ever had. The city of Lausanne offers an great working environment, with the beautiful Léman Lake and mountains nearby. Additionally, both the city and the EPFL campus are modern and dynamic, making it easy to meet new people. I met so many individuals from different countries, which allowed me to practice my english daily and learn more about various cultures. Moreover, being based in Lausanne gave me the opportunity to visit many wonderful places in Switzerland, including scientific sites of interest such as CERN in Geneva and the Einstein Museum in Bern. I am deeply grateful for this experience.
