ECN publicatie:
Neuro-fuzzy control in a steam boiler
Gepubliceerd door: Publicatie datum:
ECN 1998
ECN publicatienummer: Publicatie type:
ECN-RX--98-030 Artikel wetenschap tijdschrift
Aantal pagina's:

Gepubliceerd in: Paper, presented at the 6th European congress on intelligent techniques and soft computing (EUFIT'98), Aachen, Germany, Septembe (), , , Vol., p.-.

A neuro-fuzzy controller has been designed for a burner system in anindustrial gas-fuelled steam boiler that maximises energy efficiency under the constraint of a maximum NOx emission level. It can act as a feedforward precompensator, generating nearly optimal control actions even under strong load changes or large variations in fuel gas composition. The fuzzy network of the controller was trained on real historical process data. The underlying assumption is that the non-linear static map generated by the neural network can adequately represent the system's inverse dynamic behaviour in the ranges of interest. In this controller design strategy, process modelling is not necessary. The controllability was hence improved, resulting in a higher boiler energy efficiency and lower electrical energy consumption by the recycle flow ventilator, while the NOx emission limit is not exceeded. 4 refs.

Terug naar overzicht.