Model-based RWM Identification, ELM-Filtering, and RWM Control on the DIII-D Tokamak

Resistive-wall-mode (RWM) modeling helps not only to identify the RWM mode, but also to optimize model-based control. Based on modeled sensor signals, a static matched filter is constructed to identify the RWM mode. Using a ``picture frame" wall model, a dynamic Kalman filter has been developed to discriminate edge localized modes (ELMs) from RWMs. From this ELM-filtered RWM mode identification, an enhanced RWM feedback controller has been implemented for \hbox{DIII-D} plasmas. Recent experiments showed the effectiveness of the Kalman filter scheme; the feedback coils were rarely excited during ELMs, while responding to RWMs. To investigate proper RWM responses, an optimized Kalman filter parameter set has been found and evaluated. Meanwhile, using a thin wall treatment, an eigenmode approach has been adopted to represent the vessel more accurately.