Comparison of start-up runaway electron generation simulations using the SCENPLINT code with JET experimental observations

Abstract This paper explores the process of runaway electron (RE) formation during the start-up of a tokamak discharge. This has been done by simulating the process of RE generation and, importantly their losses, self-consistently with the behavior of many other discharge parameters, using the SCENPLINT code, which has been updated to include RE physics. For the first time, an attempt is made to compare such simulations with experimental observations on start-up RE formation at JET. This proved to be a difficult task due to the many parameters that are self-consistently calculated by the code, the possible explosive growth of RE and the fact that the diagnosis of the simulated parameters is not always very accurate. A sensitivity study on the impact of various simulation assumptions was conducted. It was found that because during the JET start-up the ratio of the electric field to the critical electric field, E / E c is limited, using the secondary generation model proposed by Aleynikov and Breizman ( Phys. Rev. Lett. 114 155001) ensures a better match to the experimental observations. This model predicts a reduced secondary generation when E / E c < 5 for start-up plasmas. Furthermore, the simulation results depended strongly on the assumed level of RE losses. Assuming the RE confinement time scales with the discharge current provide a better result. It was not, however, possible to determine which RE loss model, one based on drift-orbit losses or another based on RE diffusion due to magnetic turbulence, provided a better match to the experimental observations. These findings will be crucial to improve predictions of start-up RE formation in future devices such as ITER.