A Kalman Filter for n > 1 Resistive-Wall-Mode Identification and Feedback Control Modeling

For high performance tokamak plasmas, noise from the edge-localized modes (ELMs) is prevalent. In order to provide feedback control, ELM noise must be discriminated from that of resistive wall modes (RWMs). Although a Kalman filter has been developed that discriminates the ELM noise from the n=1 RWM, no Kalman filter has been constructed to include RWMs for n>1. This project will develop a Kalman filter that discriminates non-RWMs (e.g., ELMs) from n>1 RWMs, in addition to the n=1 RWM. Phase I developed a state-space model that serves as the basis for constructing a Kalman filter that is compatible with n=1 and n>1 RWMs. During Phase II, a Kalman filter compatible with low-n (e.g., n = 1, 2, and 3) RWMs will be constructed, based on noise characteristics from experimental data, as well as on noise modeling. Model validation will be performed in real-time experiments, followed by more advanced controller algorithm development. Commercial Applications and Other Benefits as described by the awardee: An improved Kalman filter would provide ELM-discrimination to n>1 RWMs, as well as to the n=1 RWM. The algorithm for ELM-noise discrimination from these low-n RWM modes will be conducive to RWM feedback control.