Title :
Modeling and recursive parameter estimation of nonlinear plants
Author :
Yoshitani, Naoharu
Author_Institution :
Sch. of Sci. & Eng., Teikyo Univ., Utsonimiya, Japan
Abstract :
This paper describes a modeling procedure for nonlinear plants and proposes a recursive estimation algorithm based on the model constructed through this procedure. Here, the plant model is a kind of the Hammerstein model, where a nonlinear static model is followed by a linear dynamic model in series connection. Many real plants can be represented by this structure. Parameters of both models are recursively estimated with the proposed algorithm named as “parallel and recursive estimation of static and dynamic characteristics”. The estimations of both static and dynamic characteristics are performed in parallel, by using the static and the dynamic model respectively. In the algorithm, mutual interference between the both estimations is minimized to achieve high estimation performance. The algorithm has been confirmed effective by numerical simulations. Its main idea has been applied to real commercial plants successfully
Keywords :
nonlinear systems; recursive estimation; Hammerstein model; dynamic characteristics; linear dynamic model; modeling; mutual interference; nonlinear plants; nonlinear static model; parallel estimation; recursive parameter estimation; static characteristics; Fluctuations; Industrial plants; Interference; Monitoring; Neural networks; Parameter estimation; Production; Recursive estimation; Sampling methods; Steady-state;
Conference_Titel :
Control Applications, 1997., Proceedings of the 1997 IEEE International Conference on
Conference_Location :
Hartford, CT
Print_ISBN :
0-7803-3876-6
DOI :
10.1109/CCA.1997.627587