Data-driven prediction of the: L–H transition power threshold in the EAST tokamak using ensemble learning regression
- 1 March 2026
- journal article
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 68 (3) , 035024
Abstract
Accurate prediction of the L–H transition power threshold is essential for achieving steady-state high-performance burning plasmas in future fusion reactors. In this study, we develop a machine learning–based regression framework to predict the plasma power loss across the separatrix, P loss , at the onset of the L–H transition. The model is trained on 3308 EAST discharges; from each discharge, we extract exactly one sample at the onset of its first L–H transition. Three ensemble learning models—random forest, XGBoost, and CatBoost—are systematically compared. As a benchmark, we compare our results with the empirical ITPA scaling law, which was derived from the high-density branch of the L–H power threshold in favorable ion ∇ B drift configurations of ITER-like plasmas. On our present EAST database, which spans both high-density and low-density regimes and includes discharges with both favorable and unfavorable ∇ B drift directions, the CatBoost model attains the highest predictive accuracy and yields significantly lower errors than the multivariate linear regression baseline. Model interpretation is performed using SHAP (SHapley Additive exPlanations). The SHAP rankings indicate that the key variables appearing in empirical L–H scaling relations ( S , n ¯ e , and B T ) remain the dominant predictors, while the plasma current I p is consistently identified as an additional leading contributor. Equilibrium-related and geometry-related parameters ( δ top , δ bot , q 95 , κ ) also exhibit non-negligible contributions. Underexplored spatial variables such as Z LX show a notable impact on the power threshold within the present EAST database. These results suggest that the machine-learning analysis can reproduce the main dependencies embodied in conventional scaling laws, while also indicating additional effects that are not explicitly encoded in the Martin scaling. The present framework provides an alternative data-driven approach for predicting the L–H transition, which may be particularly useful for exploring how multiple coupled control parameters jointly influence the L–H power threshold in complex, reactor-relevant operating scenarios.Keywords
Funding Information
- EAST team
- Chuzhou University Initial Funding (2024qd23)
- National Key R&D Program of China (2022YFE03020004)
- National Natural Science Foundation of China (11275308)
This publication has 45 references indexed in Scilit:
- Safety factor and turbulence dynamics dependence of the L-H power threshold on DIII-DPhysics of Plasmas, 2019
- Automatic detection of L-H transition in KSTAR by support vector machineFusion Engineering and Design, 2018
- L–H power threshold studies with tungsten/carbon divertor on the EAST tokamakRadiation Effects and Defects in Solids, 2016
- Characteristics of dithering cycles during the L–I–H transition on Experimental Advanced Superconducting Tokamak (EAST)Physics Letters A, 2013
- L–H transition studies on DIII-D to determine H-mode access for operational scenarios in ITERNuclear Fusion, 2011
- Machine learning for the identification of scaling laws and dynamical systems directly from data in fusionNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2010
- Support vector machine-based feature extractor for L/H transitions in JETReview of Scientific Instruments, 2010
- The H-mode power threshold in hydrogen plasmas in DIII-DNuclear Fusion, 2009
- H mode power threshold database for ITERNuclear Fusion, 1996
- Experimental evidence for neoclassical ion transport effects in the H-transition of ASDEXNuclear Fusion, 1985