Summary: ECAPT'94: PORTO 2426 March 1994, Third Meeting of BRITEEURAM Concerted Action on Process
VisionGuided Flame Control Using Fuzzy Logic
and Neural Networks \Lambda
Wenjing Tao and Hans Burkhardt
Technische Informatik I
Technische Universit¨at HamburgHarburg
21071 Hamburg, Germany
This paper describes a visionguided closed loop control for stationary luminous
flames using fuzzy logic and neural networks. Experts' knowledge and training
patterns can be incorporated into fuzzy rules, which are represented in the form of
neurons. Based on the learning capability of the neural networks, the system can
adjust its performance automatically.
Combustion processes are very complex reactive fluid dynamical processes. In order
to optimize these processes we need a large amount of information derived from sensors.
In general one kind of sensor (physical, electrochemical, optical, etc.) alone is not suf
ficient to cover the overall characteristics for analysis and ultimately for the control of