Abstract
Seven different methods for online system identification are discussed and compared for the realistic case when the input-output data are contaminated with noise. The performance of each method is evaluated for different noise-to-signal ratios, for a simulated 2nd-order system, as well as for data obtained from a 2-stage heat-exchanger system under operating conditions. The comparison provides considerable insight into the application of these methods to the online identification of actual processes.

This publication has 0 references indexed in Scilit: