Statistical analysis of plasma shape influence on the power threshold to access the H-mode

In the vast majority of tokamaks, the high confinement mode (H mode) has been systematically achieved and studied. The transition to the H-mode is a very complex phenomenon but it can be represented by the relation between a limited number of macroscopic quantities. The predicted power ( P Thresh ) for accessing the H-mode is generally estimated by statistical analysis of multi-machine experimental data and typically presented in terms of equations in power law monomial form. Even if the positive effects of the plasma shape on the confinement time have been clearly documented, the dependence of the P Thresh on elongation and triangularity is still an aspect no completely clarified. The objective of this paper is to assess whether the P Thresh dependency on the shape of the plasma can be derived by a statistical approach using a recent version of ITPA International Global Threshold Data Base (IGDBTHv6b). A statistical method to select relevant macroscopic quantities in the candidate models for the scaling of P Thresh is presented and a multi-machine scaling law, using non-linear regression techniques, is derived.