Vol.8, No 1, 2010 pp. 47 - 56
UDC 66.096.5, 519.673, 681.51
ADVANCED EVOLUTIONARY OPTIMIZATION FOR INTELIGENT MODELING AND CONTROL OF FBC PROCESS
Žarko Ćojbašić1, Vlastimir Nikolić1, IvanĆirić1, Sorin Grigorescu2
1University of Niš, Faculty of Mechanical Engineering, A. Medvedeva 14, Niš, Serbia, E-mail: zcojba@ni.ac.rs
2Institute of Automation, University of Bremen, Otto-Hahn-Allee, NW1, 28359 Bremen, Germany
In this paper, the computationally intelligent modeling approach for fluidized bed combustion process is considered, as well as intelligent process control based on developed models and soft computing methodologies. Applied genetic optimization and fuzzy modeling approach are combined with proposed intelligent control strategies. Efficient fuzzy nonlinear FBC process modeling by combining several linearized models is presented, as well as fuzzy and conventional process control systems optimized by real coded genetic algorithms. The obtained results indicate that the computationally intelligent approach can be successfully applied for modeling and control of such a complex fluidized bed combustion process.
Key words: Computational Intelligence, Fluidized Bed Combustion, Fuzzy Systems, Real Coded Genetic Algorithms
NAPREDNA EVOLUCIONA OPTIMIZACIJA ZA INTELIGENTNO MODELIRANJE I UPRAVLJANJE PROCESOM SFS
U ovom radu razmatran je pristup modeliranju procesa sagorevanja u fluidizovanom sloju zasnovan na računarskoj inteligenciji, kao i inteligentno upravljanje procesom bayirano na rayvijenim modelima i metodama mekog računa. Primenjena genetska optimizacija i fazi pristup modeliranju kombinovani su sa predloženim strategijama inteligentnog upravljanja. Predstavljeno je efikasno fazi nelinearno modeliranje procesa SFS kombinovanjem linearizovanih modela, kao i realno kodirana genetska optimizacija fazi i konvencionalnog sistema upravljanja procesom. Dobijeni rezultati ukazuju da se pristup zasnovan na računarskoj inteligenciji uspešno može primenjivati na ovakav kompleksni process sagorevanja u fluidizovanom sloju.
Ključne reči: računarska inteligencija, sagorevanje u fluidizovanom sloju, fazi sistemi, realno kodirani genetski algiritmi