Estimation of plasma properties using an extended Kalman filter with plasma global models

Abstract

A physically-constrained extended Kalman filter (EKF) is applied to various zero-dimensional global models for the estimation of plasma properties using time-dependent experimental data such as the plasma density or ion flux. The capability of the EKF is demonstrated to estimate unknown system states simultaneously, such as reaction rate coefficients and the absorbed electron input power, which can be difficult, if not impossible, to measure experimentally. Global models accounting for pure argon reactions and argon-oxygen reactions are used in this work to demonstrate the ability of the filter to estimate dynamic and complex systems. The results obtained from the EKF plasma global model illustrate that model-data fusion techniques can be used to estimate plasma properties and processes for time-varying systems, such as pulsed discharges.