System Identification Methods for Industrial Plant

R.J. Treasure, U. Kruger, and J.E. Cooper (UK)


System Identification, State Space, Model Structures.


In this paper three models for system identification are compared. The performance of state-space, finite impulse response (FIR) and autoregressive exogenous (ARX) models are compared on the basis of data obtained from a simulation of a fluid catalytic cracking unit. The focus of this comparison is laid upon the number of the model parameters and the accuracy of the obtained models. The state space model uses fewer model parameters and also provides the best prediction of the output variables.

Important Links:

Go Back