Soft Sensor for Flue Gas Monitoring in Power Station Boilers

Abstract

The flue gas monitoring plays a vital role in thermal power plants as these emissions contribute to the majority of air pollution and global warming. Hence an intelligent sensor can be used in place of gas analyzer which is economical as well as provides information regarding the flue gas emissions based on the color of the furnace flame. The intelligent sensor is capable of classifying the flame images based on the various features that are extracted. The features extracted vary from complete combustion category to incomplete combustion category. The attributes are selected using the CFS method. The selected attributes are then classified by the Radial Basis Function (RBF) classifier so as to give the information about the flue gas emissions based on the color of the flame images at the furnace level itself. The simulation was done using Weka tool.