Sensitivity Analysis of Transport and Radiation in NeuralPlasmaODE for ITER Burning Plasmas

Understanding how key physical parameters influence burning plasma behavior is essential for reliable and efficient operation of ITER. In this study, we extend the NeuralPlasmaODE framework, a multi-region, multi-timescale model based on neural ordinary differential equations (Neural ODEs), to perform a systematic sensitivity analysis of transport and radiation mechanisms in ITER plasmas. Normalized sensitivities of core and edge densities and temperatures are computed with respect to transport diffusivities, electron cyclotron radiation (ECR) parameters, impurity fractions, and ion orbit loss (IOL) timescales. The analysis is based on a trained model for ITER inductive scenario 2. Results highlight the dominant role of magnetic field strength, safety factor, and impurity content in shaping energy transport and confinement. We also show how nonlinear coupling between thermal diffusivity and temperature gradients induces a self-regulating mechanism that mitigates thermal runaway. While ECR and IOL parameters have weaker influence overall, they can still affect edge power balance in certain conditions. This work supports robust scenario design and illustrates the extensibility of NeuralPlasmaODE for predictive modeling and control-oriented fusion applications.