Estimation of plasma equilibrium parameters via a neural network approach

Plasma equilibrium parameters such as position, X-point, internal inductance, and poloidal beta are essential information for efficient and safe operation of tokamak. In this work, the artificial neural network is used to establish a non-linear relationship between the measured diagnostic signals and selected equilibrium parameters. The estimation process is split into a preliminary classification of the kind of equilibrium (limiter or divertor) and subsequent inference of the equilibrium parameters. The training and testing datasets are generated by the tokamak simulation code (TSC), which has been bench-marked with the EAST experimental data. The noise immunity of the inference model is tested. Adding noise to model inputs during training process is proved to have a certain ability for maintaining performance.