Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Combustion optimization using an expert system and neural networks

Conference ·
OSTI ID:427871
; ; ; ; ;  [1];  [2]
  1. Lehigh Univ., Bethlehem, PA (United States). Energy Research Center
  2. Potomac Electric Power Co., Upper Marlboro, MD (United States)
Lehigh University`s Energy Research Center and the Potomac Electric Power Company have been developing software for use by plant personnel in turning a pulverized coal-fired boiler to achieve optimized combustion. The software is based on the interactions of an expert system, neural network, and a mathematical optimization algorithm. It uses the expert system to safely guide the plant engineer through a series of parametric boiler tests and gather a data base which characterizes boiler operation over a wide range of conditions. The neural network portion develops non-linear mapping functions between the outputs of NO{sub x}, heat rate, LOI, opacity, and the controllable boiler input parameters. These mapping functions are then analyzed by the mathematical optimization algorithm and the optimal boiler operating conditions are identified. This paper describes the application of the software to corner-fired boilers with either conventional or low NO{sub x} burners and overfire air.
OSTI ID:
427871
Report Number(s):
CONF-961006--; ISBN 0-7918-1795-4
Country of Publication:
United States
Language:
English