Spectral matching for parameter estimation in nonlinear input-output models

Author(s):  
O. Markusson ◽  
H. Hjalmarsson
Author(s):  
Héctor Botero ◽  
Hernán Álvarez

This paper proposes a new composite observer capable of estimating the states and unknown (or changing) parameters of a chemical process, using some input-output measurements, the phenomenological based model and other available knowledge about the process. The proposed composite observer contains a classic observer (CO) to estimate the state variables, an observer-based estimator (OBE) to obtain the actual values of the unknown or changing parameters needed to tune the CO, and an asymptotic observer (AO) to estimate the states needed as input to the OBE. The proposed structure was applied to a CSTR model with three state variables. With the proposed structure, the concentration of reactants and other CSTR parameters can be estimated on-line if the reactor and jacket temperatures are known. The procedure for the design of the proposed structure is simple and guarantees observer convergence. In addition, the convergence speed of state and parameter estimation can be adjusted independently.


Econometrica ◽  
1973 ◽  
Vol 41 (6) ◽  
pp. 1167 ◽  
Author(s):  
I. W. Sandberg
Keyword(s):  

1994 ◽  
Vol 37 (3) ◽  
pp. 345-356 ◽  
Author(s):  
Jeng-Ming Chen ◽  
Bor-Sen Chen ◽  
Wei-Sheng Chang

Sign in / Sign up

Export Citation Format

Share Document