Vol.1, No 10, 2003 pp. 1393- 1406
UDC 662.61
INTELLIGENT CONTROL OF COMPLEX
COMBUSTION PROCESSES*
Žarko Ćojbašić1, Volker Stephan2,
Horst-Michael Gross2, Jürgen Wernstedt2
1University of Niš, Faculty of Mechanical
Engineering, Serbia and Montenegro
2Ilmenau Technical University, Faculty
of Computer Science and Automation, Germany
Abstract. In this paper intelligent control approaches for complex
combustion processes have been considered. Successful applications of computational
intelligence for advanced modelling, identification and control of thermal
plants are presented. Specifically, neuro-fuzzy modelling of fluidized
bed combustion process, and also intelligent process control based on developed
models have been considered. Also, we have presented a novel control scheme
for an industrial hard-coal combustion process in a power plant based on
reinforcement-learning in combination with neural networks. These intelligent
control approaches are aimed at complying with ever stricter requirements
for environmental protection while maximizing the efficiency factor simultaneously
keeping other process parameters within predefined limits.
Key words: Intelligent Control, Combustion Processes, Fuzzy
Systems, Neural Networks, Reinforcement Learning, Genetic Algorithms
INTELIGENTNO UPRAVLJANJE SLOŽENIM
PROCESIMA SAGOREVANJA
U radu su razmatrani inteligentni pristupi upravljanju složenim procesima
sagorevanja. Razvijene su uspešne primene računarske inteligencije za modeliranje,
identifikaciju i upravljanje termoenergetskim postrojenjima. Konkretno,
razmatran je neuro-fazi pristup modeliranju procesa sagorevanja u fluidizovanom
sloju kao i inteligentno procesno upravljanje na bazi razvijenih modela.
Osim toga projektovana je i nova upravljačka šema za industrijski proces
sagorevanja uglja u termoelektrani, zasnovana na obučavanju sa osnaživanjem
u kombinaciji sa veštačkim neuronskim mrežama. Ovi inteligentni upravljački
pristupi su usmereni na zadovoljavanje sve strožijih zahteva u pogledu
zaštite okoline uz istovremenu maksimizaciju faktora efikasnosti i uz simultano
održavanje drugih parametara procesa u definisanim granicama.
Ključne reči: inteligentno upravljanje, procesi sagorevanja,
fazi sistemi, neuronske mreže, obučavanje sa osnaživanjem, genetski algoritmi
