Data-Constrained Reduced ELM Model for Integrated Tokamak Transport Simulations

Y. Tao, K. Shabbir, T. Rafiq, E. Schuster

68th Division of Plasma Physics (DPP) Annual Meeting of the American Physical Society (APS)

Chicago, IL, USA, November 2-6, 2026

A data-constrained, reduced edge-localized-mode (ELM) model has been developed for control-oriented integrated tokamak transport simulations in COTSIM. The framework provides a modular description of ELM triggering, crash, and recovery while supporting multiple pedestal closures. ELM onset is determined using reduced ballooning and peeling criteria based on the evolving pressure gradient, magnetic shear, plasma shaping, and edge current [1]. The crash is modeled as a rapid, radially localized enhancement of electron and ion heat transport. Three recovery formulations are introduced to ensure compatibility with different pedestal closures: natural pedestal rebuilding through the transport equations for transport-based pedestal models, and recovery through a dynamic recovery variable or a reduced pedestal energy balance model when coupled to predictive pedestal models such as PEDESTAL [2], which provide the target pedestal state. Bayesian inference is employed to calibrate model parameters against time-resolved ELM diagnostics and quantify their uncertainties. The resulting framework enables computationally efficient, experimentally constrained ELM simulations suitable for integrated transport modeling, scenario development, and plasma control studies.

[1] H. R. Wilson et al., Nuclear Fusion 40, 713 (2000).
[2] T. Onjun et al., Physics of Plasmas 9, 5018 (2002).

*Supported by the U.S. DOE under Awards DE-SC0010661.