Fuzzy model of inductively coupled plasma by adaptive-network-based fuzzy inference system (ANFIS)

摘要

[[abstract]]The chemical and physical reactions of plasma are complex and nonlinear. The controller design is limited at some operation points by the conventional method. For a nonlinear controller design, to build a model will be beneficial. This paper reports a fuzzy model that simulates the behaviors of inductive mode and capacitive mode in inductively coupled plasmas by using an adaptive-network-based fuzzy inference system (ANFIS). In this study, 13.56 MHz RF power and pressure are input variables and electron density is output variable. The electron density was measured by 36 GHz heterodyne interferometer. The training data for ANFIS were collected by varying different RF power at different pressure levels. This research has demonstrated a simplified method to synthesize a nonlinear system that can cover the characteristics of capacitive mode and inductive mode. It will be helpful to simulate a controller performance before applying to a real plasma system.