Input-state estimation of a spatially distributed tubular gasification reactor

Abstract

Abstract The problem of on-line estimating the gas feed flow input and the 13 (2 flows, 1 temperature and 10 component)-profile state of a bistable tubular gasification reactor is addressed. The estimator must be developed from the underlying first principles nonlinear partial differential equation (PDE) model in conjunction with feed and sensitive-location temperature measurements. The problem is solved through the dynamical inversion of an efficient finite differences (FD)-based differential-algebraic equation (DAE) model approximation of the PDE one. It is found that, using a model previously calibrated against experimental data, the problem can be solved with a DA inverse made of 60 ODEs and 180 AEs.