Simple method for parameter identification of a nonlinear Greitzer compressor model

Abstract

This work introduces a simple method to estimate state variables and identify parameters of a nonlinear dynamic Greitzer compressor model. The observer is based upon an extended Kalman filter, which estimates the dynamic states as well as a subset of parameters. In a Monte-Carlo-fashioned approach, the remaining set of parameters is then identified by minimizing an objective function representing the error between the measured variables and their estimates. The developments are demonstrated in numerical simulations.