Design of fuzzy logic controller for combustion control of stoker-fired boilers based on behavior modeling

This paper uses a behavior modeling approach to design a fuzzy logic controller for combustion control of stoker-fired boilers. Its idea is, according to recorded historical data, to model a boiler combustion system by defining types of dynamic behavior. On the basis of behavior classification, design of a neuro-fuzzy control system is performed in two steps. First we model a boiler combustion process with unknown structure by choosing a set of simplified systems with equivalent behavior as "templates" to optimize their fuzzy controllers off-line. Second, we identify a dynamic behavior and fire its correspond fuzzy logic controller by matching the "templates" online. This control system is applied into several industrial stoker-fired boilers in different places.