Title :
Neural networks with multiple-state neurons for nitrogen oxide (NO x) emissions modeling and advisory control
Author :
Reinschmidt, Kenneth F. ; Ling, Bo
Author_Institution :
Stone & Webster Adv. Syst. Dev. Services Inc., Boston, MA, USA
fDate :
27 Jun- 2 Jul 1994
Abstract :
In this paper, neural networks are developed for the modeling and control of the nitrogen oxide (NOx) emissions from coal-fired boilers. Both the neural network NOx simulation model and the neural network controller are trained with real plant data. Multiple-state neurons are used in the neural network controller to restrict the range of the controller. A modified back propagation algorithm is used in the network training. The results indicate that the feedforward neural networks can be used as a real-time advisor for the plant operator to determine the optimum settings for lower NOx emissions
Keywords :
air pollution; air pollution control; air pollution measurement; backpropagation; boilers; control engineering computing; decision support systems; feedforward neural nets; nitrogen compounds; optimal control; NO; NO2; NOx emissions; advisory control; coal-fired boilers; feedforward neural networks; modified back propagation algorithm; multiple-state neurons; neural network controller; nitrogen oxide emissions modeling; Boilers; Combustion; Control systems; Neural networks; Neurons; Nitrogen; Nonlinear control systems; Power system modeling; Response surface methodology; Temperature control;
Conference_Titel :
Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-1901-X
DOI :
10.1109/ICNN.1994.374822
