Intelligent Flue Gas Monitoring in Power Station Boilers

This research work which deals with monitoring of flue gas emissions is an intelligent technique for flue gas monitoring based on the colour of the flame images acquired from the furnace will be discussed here. A combination of image processing algorithms with Bayesian and intelligent classifiers are used to identify the flue gas emissions in order to ensure complete combustion. The features are extracted with Image J and reduced using Support Vector Machine (SVM). The classification of the flame images was obtained from the selected features using both the intelligent and Bayes classifiers. The intelligent classifiers were found to be beneficial compared to the Bayes classifier to monitor the flue gas emissions at the furnace level thereby providing a feed forward control action.